PR NousResearch#13743 replaced the global MAX_TEXT_LENGTH=4000 with a per-provider
table and a user-override 'max_text_length:' key, but the user-guide
TTS page documented no length behaviour at all. Users hitting truncation
had no way to discover the new caps or the override.

Add an 'Input length limits' subsection after the existing Configuration
YAML block: provider default caps (Edge 5000 / OpenAI 4096 / xAI 15000 /
MiniMax 10000 / Mistral 4000 / Gemini 5000 / ElevenLabs model-aware /
NeuTTS,KittenTTS 2000), ElevenLabs model_id -> cap table (5k-40k), an
override example, and the validation rules (non-positive / non-integer /
boolean values fall through to the provider default).

* fix(tui): raise picker selection contrast with inverse + bold

Selected rows in the model/session/skills pickers and approval/clarify
prompts only changed from dim gray to cornsilk, which reads as low
contrast on lighter themes and LCDs (reported during TUI v2 blitz).

Switch the selected row to `inverse bold` with the brand accent color
across modelPicker, sessionPicker, skillsHub, and prompts so the
highlight is terminal-portable and unambiguous. Unselected rows stay
dim. Also extends the sessionPicker middle meta column (which was
always dim) to inherit the row's selection state.

* fix(model-switch): drop stale provider from fallback chain and env after /model

Reported during the TUI v2 blitz test: switching from openrouter to
anthropic via `/model <name> --provider anthropic` appeared to succeed,
but the next turn kept hitting openrouter — the provider the user was
deliberately moving away from.

Two gaps caused this:

1. `Agent.switch_model` reset `_fallback_activated` / `_fallback_index`
   but left `_fallback_chain` intact. The chain was seeded from
   `fallback_providers:` at agent init for the *original* primary, so
   when the new primary returned 401 (invalid/expired Anthropic key),
   `_try_activate_fallback()` picked the old provider back up without
   informing the user. Prune entries matching either the old primary
   (user is moving away) or the new primary (redundant) whenever the
   primary provider actually changes.

2. `_apply_model_switch` persisted `HERMES_MODEL` but never updated
   `HERMES_INFERENCE_PROVIDER`. Any ambient re-resolution of the runtime
   (credential pool refresh, compressor rebuild, aux clients) falls
   through to that env var in `resolve_requested_provider`, so it kept
   reporting the original provider even after an in-memory switch.

Adds three regression tests: fallback-chain prune on primary change,
no-op on same-provider model swap, and env-var sync on explicit switch.

* fix(tui): @folder: only yields directories, @file: only yields files

Reported during TUI v2 blitz testing: typing `@folder:` in the composer
pulled up .dockerignore, .env, .gitignore, and every other file in the
cwd alongside the actual directories. The completion loop yielded every
entry regardless of the explicit prefix and auto-rewrote each completion
to @file: vs @folder: based on is_dir — defeating the user's choice.

Also fixed a pre-existing adjacent bug: a bare `@file:` or `@folder:`
(no path) used expanded=="." as both search_dir AND match_prefix,
filtering the list to dotfiles only. When expanded is empty or ".",
search in cwd with no prefix filter.

- want_dir = prefix == "@folder:" drives an explicit is_dir filter
- preserve the typed prefix in completion text instead of rewriting
- three regression tests cover: folder-only, file-only, and the bare-
  prefix case where completions keep the `@folder:` prefix

* fix(tui): truncate long picker rows so the height stays stable

A6 added a fixed-height grid (Array.from({length: VISIBLE})), but the
row <Text> itself had no wrap prop so Ink defaulted to wrap="wrap".
A sufficiently long model or provider name would wrap to a second
visual line and bounce the overall picker height right back — which
is exactly what reappeared during the TUI v2 blitz retest on /model.

Pin every picker row (and the empty-state / padding rows) to
wrap="truncate-end" so each slot is guaranteed one line.  Applies
across modelPicker, sessionPicker, and skillsHub.

* fix(tui): stabilize slash-completion dropdown height

The completion popup (e.g. typing `/model`) grew from 8 rows at
compIdx=0 up to 16 rows at compIdx≥8 — the slice end was `compIdx + 8`
so every arrow-down added another rendered row until the window filled.
Reported during TUI v2 retest: "as i scroll and more options appear,
for some reason more options appear and it expands the height".

Fixed viewport (`COMPLETION_WINDOW = 16`) centered on compIdx, clamped
so it never slides past the array bounds.  Renders exactly
`min(WINDOW, completions.length)` rows every frame.

* fix(tui): pager supports scrolling (up/down/page/top/bottom)

The pager overlay backing /history, /toolsets, /help and any paged slash
output only advanced with Enter/Space and closed at the end. Could not
scroll back, scroll line-by-line, or jump to endpoints.

Adds Up/Down (↑↓, j/k), PgUp (b), g/G for top/bottom, keeps existing
Enter/Space/PgDn forward-and-auto-close, and clamps offset so
over-scrolling past the last page is a no-op.

* fix(tui): preserve prior segment output on Ctrl+C interrupt

interruptTurn only flushed the in-flight streaming chunk (bufRef) to
the transcript before calling idle(), which wiped segmentMessages and
pendingSegmentTools. Every tool call and commentary line the agent had
already emitted in the current turn disappeared the moment the user
cancelled, even though that output is exactly what they want to keep
when they hit Ctrl+C (quote from the blitz feedback: "everything was
fine up until the point where you wanted to push to main").

Append each flushed segment message to the transcript first, then
render the in-flight partial with the `*[interrupted]*` marker and its
pendingSegmentTools. Sys-level "interrupted" note still fires when
there is nothing to preserve.

* fix(tui): route skills.manage through the long-handler thread pool

`/skills browse` is documented to scan 6 sources and take ~15s, but the
gateway dispatched `skills.manage` on the main RPC thread.  While it
ran, every other inbound RPC — completions, new slash commands, even
`approval.respond` — blocked until the HTTP fetches finished, making
the whole TUI feel frozen.  Reported during TUI v2 retest:
"/skills browse blocks everything else".

`_LONG_HANDLERS` already exists precisely for this pattern (slash.exec,
shell.exec, session.resume, etc. run on `_pool`).  Add `skills.manage`
to that set so browse/search/install run off the dispatcher; the fast
`list` / `inspect` actions pay a negligible thread-pool hop.

* improve llama.cpp skill

* fix(skills/llama-cpp): concise description, restore python bindings, fix curl

- Description truncated to 60 chars in system prompt (extract_skill_description),
  so the 500-char HF workflow description never reached the agent; shortened to
  'llama.cpp local GGUF inference + HF Hub model discovery.' (56 chars).
- Restore llama-cpp-python section (basic, chat+stream, embeddings,
  Llama.from_pretrained) and frontmatter dependencies entry.
- Fix broken 'Authorization: Bearer ***' curl line (missing closing quote;
  llama-server doesn't require auth by default).

* fix(gateway): always inject reply-to pointer, not just when quoted text is absent (#13676)

The [Replying to: "..."] prefix is disambiguation, not deduplication. When
a user explicitly replies to a prior message, the agent needs a pointer to
which specific message they're referencing — even when the quoted text
already exists somewhere in history. History can contain the same or
similar text multiple times; without an explicit pointer the agent has to
guess (or answer for both subjects), and the reply signal is silently
dropped.

Example: in a conversation comparing Japan and Italy, replying to the
"Japan is great for culture..." message and asking "What's the best time
to go?" — previously the found_in_history check suppressed the prefix
because the quoted text was already in history, leaving the agent to
guess which destination the user meant. Now the pointer is always present.

Drops the found_in_history guard added in #1594. Token overhead is
minimal (snippet capped at 500 chars on the new user turn; cached prefix
unaffected). Behavior becomes deterministic: reply sent ⇒ pointer present.

Thanks to smartyi for flagging this.

* feat(image-gen): add GPT Image 2 to FAL catalog (#13677)

Adds OpenAI's new GPT Image 2 model via FAL.ai, selectable through
`hermes tools` → Image Generation. SOTA text rendering (including CJK)
and world-aware photorealism.

- FAL_MODELS entry with image_size_preset style
- 4:3 presets on all aspect ratios — 16:9 (1024x576) falls below
  GPT-Image-2's 655,360 min-pixel floor and would be rejected
- quality pinned to medium (same rule as gpt-image-1.5) for
  predictable Nous Portal billing
- BYOK (openai_api_key) deliberately omitted from supports so all
  users stay on shared FAL billing
- 6 new tests covering preset mapping, quality pinning, and
  supports-whitelist integrity
- Docs table + aspect-ratio map updated

Live-tested end-to-end: 39.9s cold request, clean 1024x768 PNG

* refactor(delegate): drop dead default_toolsets from CLI default config

delegation.default_toolsets was declared in cli.py's CLI_CONFIG default
dict and documented in cli-config.yaml.example, but never read: none of
tools/delegate_tool.py, _load_config(), or any call site ever looked it
up. The live fallback is the DEFAULT_TOOLSETS module constant at
tools/delegate_tool.py:101, which stays as-is.

hermes_cli/config.py's DEFAULT_CONFIG["delegation"] already omits the
key — this commit aligns cli.py with that.

Adds a regression test in tests/hermes_cli/test_config_drift.py so a
future refactor that re-adds the key without wiring it up to
_load_config() fails loudly.

Part of Initiative 2 / M0.5.

* docs(delegate): remove default_toolsets from example config and docs

Matches the default-config removal in the preceding commit.
default_toolsets was documented for users to set but was never actually
read at runtime, so showing it in the example config and the delegation
user guide was misleading.

No deprecation note is added: the key was always a no-op, so users who
copied it from the example continue to see no behavior change. Their
config.yaml still parses; the key is just silently unused, same as
before.

Part of Initiative 2 / M0.5.

* test(delegate): make default_toolsets regression test robust to user config

The prior form of this test asserted on CLI_CONFIG["delegation"] after
importing cli, which only passed by accident of pytest-xdist worker
scheduling. cli._hermes_home is frozen at module import time (cli.py:76),
before the tests/conftest.py autouse HERMES_HOME-isolation fixture can
fire, so CLI_CONFIG ends up populated by deep-merging the contributor's
actual ~/.hermes/config.yaml over the defaults (cli.py:359-366). Any
contributor (like me) who still has the legacy key set in their own
config causes a false failure the moment another test file in the same
xdist worker imports cli at module level.

Asserting on the source of load_cli_config() instead sidesteps all of
that: the test now checks the defaults literal directly and is
independent of user config, HERMES_HOME, import order, and worker
scheduling.

Demonstrated failure mode before this fix:
  pytest tests/hermes_cli/test_config_drift.py \
         tests/hermes_cli/test_skills_hub.py -o addopts=""
  -> FAILED (CLI_CONFIG["delegation"] contained "default_toolsets"
     from the user's ~/.hermes/config.yaml)

Part of Initiative 2 / M0.5.

* feat(gateway): recognize .pdf in MEDIA: tag extraction (#13683)

PDFs emitted by tools (report generators, document exporters, etc.) now
deliver as native attachments when wrapped in MEDIA: — same as images,
audio, and video.

Bare .pdf paths are intentionally NOT added to extract_local_files(), so
the agent can still reference PDFs in text without auto-sending them.

* fix(tui): inject VS16 so text-default emoji render as color glyphs

Models frequently emit bare codepoints like U+26A0 (⚠), U+2139 (ℹ),
U+2764 (❤), U+2714 (✔), U+2600 (☀), U+263A (☺) which, per Unicode, have
Emoji_Presentation=No and render as monochrome text-style glyphs in
terminals unless followed by VS16 (U+FE0F). Agent output leaked through
the TUI like `⚠ careful` instead of `⚠️ careful`.

Added `ensureEmojiPresentation` (lib/emoji.ts): scans for the curated
set of text-default codepoints and appends VS16 when the next char is
not already VS16, ZWJ, or a keycap-enclosing mark. Idempotent and
fast-pathed by a Unicode-range regex so ASCII-heavy text is untouched.

Applied once at the top of `Md`'s line parse. Hermes-ink's stringWidth
already accounts for VS16, so cursor/layout stays correct.

* feat(delegate): orchestrator role and configurable spawn depth (default flat)

Adds role='leaf'|'orchestrator' to delegate_task. With max_spawn_depth>=2,
an orchestrator child retains the 'delegation' toolset and can spawn its
own workers; leaf children cannot delegate further (identical to today).

Default posture is flat — max_spawn_depth=1 means a depth-0 parent's
children land at the depth-1 floor and orchestrator role silently
degrades to leaf. Users opt into nested delegation by raising
max_spawn_depth to 2 or 3 in config.yaml.

Also threads acp_command/acp_args through the main agent loop's delegate
dispatch (previously silently dropped in the schema) via a new
_dispatch_delegate_task helper, and adds a DelegateEvent enum with
legacy-string back-compat for gateway/ACP/CLI progress consumers.

Config (hermes_cli/config.py defaults):
  delegation.max_concurrent_children: 3   # floor-only, no upper cap
  delegation.max_spawn_depth: 1           # 1=flat (default), 2-3 unlock nested
  delegation.orchestrator_enabled: true   # global kill switch

Salvaged from @pefontana's PR #11215. Overrides vs. the original PR:
concurrency stays at 3 (PR bumped to 5 + cap 8 — we keep the floor only,
no hard ceiling); max_spawn_depth defaults to 1 (PR defaulted to 2 which
silently enabled one level of orchestration for every user).

Co-authored-by: pefontana <fontana.pedro93@gmail.com>

* fix(auxiliary): refresh Nous runtime credentials after aux 401s

* docs(delegate): clarify that the parent agent, not the user, populates goal/context (#13698)

The 'subagents know nothing' warning and the 'no conversation history'
constraint both said the user provides the goal/context fields. In
practice the LLM parent agent calls delegate_task; the user configures
the feature but doesn't write delegation calls. Rewording to point at
the parent agent matches how the tool actually works.

* fix(vision): resolve Nous vision model correctly in auto-detect path

Two changes:
1. _PROVIDER_VISION_MODELS: add 'nous' -> 'xiaomi/mimo-v2-omni' entry
   so the vision auto-detect chain picks the correct multimodal model.

2. resolve_provider_client: detect when the requested model is a vision
   model (from _PROVIDER_VISION_MODELS or known vision model names) and
   pass vision=True to _try_nous().  Previously, _try_nous() was always
   called without vision=True in resolve_provider_client(), causing it to
   return the default text model (gemini-3-flash-preview or mimo-v2-pro)
   instead of the vision-capable mimo-v2-omni.

The _try_nous() function already handled free-tier vision correctly, but
the resolve_provider_client() path (used by the auto-detect vision chain)
never signaled that a vision task was in progress.

Verified: xiaomi/mimo-v2-omni returns HTTP 200 with image inputs on Nous
inference API. google/gemini-3-flash-preview returns 404 with images.

* chore(release): add Ifkellx to AUTHOR_MAP for PR #12687

* fix(security): TUI approval overlay accepts blind keystrokes, CLI thread-local callback invisible to agent

Two bugs that allow dangerous commands to execute without informed user consent.

TUI (Ink): useInputHandlers consumes the isBlocked return path, but Ink's
EventEmitter delivers keystrokes to ALL registered useInput listeners. The
ApprovalPrompt component receives arrow keys, number keys, and Enter even
though the overlay appears frozen. The user sees no visual feedback, but
keystrokes are processed — allowing blind approval, session-wide auto-approve
(choice "session"), or permanent allowlist writes (choice "always") without
the user knowing.

Discovered while replicating #13618 (TUI approval overlay freezes terminal).

Fix: in useInputHandlers, when overlay.approval/clarify/confirm is active,
only intercept Ctrl+C. All other keys pass through. This makes the overlay
visually responsive so the user can see what they are selecting.

CLI (prompt_toolkit): _callback_tls in terminal_tool.py is threading.local().
set_approval_callback() is called in the main thread during run(), but the
agent executes in a background thread. _get_approval_callback() returns None
in the agent thread, falling back to stdin input() which prompt_toolkit
blocks. The user sees the approval text but cannot respond — the terminal is
unusable until the 60s timeout expires with a default "deny".

Fix: set callbacks inside run_agent() (the thread target), matching the
pattern already used by acp_adapter/server.py. Clear on thread exit to avoid
stale references.

Closes #13618

* test(approval): regression guards for thread-local callback contract

Two unit tests that pin down the threading.local semantics the CLI freeze
fix (#13617 / #13618) relies on:

- main-thread registration must be invisible to child threads (documents
  the underlying bug — if this ever starts passing visible, ACP's
  GHSA-qg5c-hvr5-hjgr race has returned)
- child-thread registration must be visible from that same thread AND
  cleared by the finally block (documents the fix pattern used by
  cli.py's run_agent closure and acp_adapter/server.py)

Pairs with the fix in the preceding commit by @Societus.

* fix(vision): route Nous main-provider vision through tier-aware backend

* fix(vision): restore tier-aware Nous vision model selection (#13703)

Revert two overreaches from #13699 that forced paid Nous vision to
xiaomi/mimo-v2-omni instead of the tier-appropriate gemini-3-flash-preview:

1. Remove "nous": "xiaomi/mimo-v2-omni" from _PROVIDER_VISION_MODELS —
   #13696 already routes nous main-provider vision through the strict
   backend, and this entry caused any direct resolve_provider_client(
   "nous", ...) aggregator-lookup path to pick the wrong model for paid.

2. Drop the 'elif vision' paid override in _try_nous() that forced
   mimo-v2-omni on every Nous vision call regardless of tier. Paid
   accounts now keep gemini-3-flash-preview for vision as well as text.

Free-tier behavior unchanged: still uses mimo-v2-omni for vision,
mimo-v2-pro for text (check_nous_free_tier() branch).

E2E verified:
  paid vision → google/gemini-3-flash-preview
  free vision → xiaomi/mimo-v2-omni
  paid text   → google/gemini-3-flash-preview
  free text   → xiaomi/mimo-v2-pro

* feat(llm-wiki): port provenance markers, source hashing, and quality signals from llm-wiki-compiler (#13700)

Three additive conventions inspired by github.com/atomicmemory/llm-wiki-compiler:

- Paragraph-level provenance: `^[raw/articles/source.md]` markers on pages synthesizing 3+ sources, so readers can trace individual claims without re-reading full source files.
- Raw source content hashing: `sha256:` in raw/ frontmatter enables re-ingest drift detection — skip unchanged sources, flag changed ones.
- Optional `confidence` and `contested` frontmatter fields let lint surface weak or disputed claims without re-reading every page's prose.

Lint gains two new checks (quality signals, source drift) and one expanded check (contradictions now surfaces frontmatter-flagged pages).

Also adds a Related Tools section pointing users who want batch/scheduled compilation at llm-wiki-compiler (Obsidian-compatible, works on the same vault).

All additions are opt-in — existing wikis need no migration. Skill version 2.0.0 -> 2.1.0.

* fix(tui): don't swallow Kimi/Qwen ~! ~? kaomoji as subscript spans

The inline markdown regex had `~([^~\s][^~]*?)~` for Pandoc-style subscript
(H~2~O, CO~2~). On models that decorate prose with kaomoji like `thing ~!`
and `cool ~?` — Kimi especially — the opener `~!` paired with the next
stray `~` on the line and dim-formatted everything between them with a
leading `_` character, mangling markdown output.

Tighten the pattern to short alphanumeric-only content (`~[A-Za-z0-9]{1,8}~`)
since real subscript never contains punctuation, spaces, or long runs.
Same tightening applied to stripInlineMarkup so width measurement stays
consistent. Classic CLI was unaffected because it renders these literally.

* refactor(tui): clean markdown.tsx per KISS/DRY

- Drop the outer no-op capture group from INLINE_RE and restructure the
  source as an ordered list of patterns-with-index-comments so each
  alternative is individually greppable. Shift group indices in MdInline
  down by one accordingly.
- Inline single-use helpers (parseFence, isFenceClose, isMarkdownFence,
  trimBareUrl) and intermediate variables (path, lang, raw, prefix, body,
  depth, task body, setext match, etc.).
- Hoist block-level regexes used inside MdImpl (FENCE_CLOSE_RE, SETEXT_RE,
  BULLET_RE, TASK_RE, NUMBERED_RE, QUOTE_RE) to top-level consts so
  they're compiled once instead of per-line.
- Collapse the duplicate compact-vs-normal blank-line branches into one
  if/!compact gap call.
- Move Fence and MdProps types to the bottom per house style.
- Shorten splitTableRow → splitRow and use optional chaining in a few
  match sites.

No behavior change; 162/162 tests pass. Net -22 LoC.

* fix(tui): /resume picker shows telegram/discord/etc sessions

Reported during TUI v2 blitz retest: /resume modal only surfaced tui/cli
rows, even though `hermes --tui --resume <id>` with a pasted telegram
session id works fine.  The handler double-fetched with explicit
`source="tui"` and `source="cli"` filters and dropped everything else on
the floor.

Drop the filter — list_sessions_rich(source=None) already excludes
child sessions (subagents, compression continuations) via its default,
and users want to resume messenger sessions from inside the TUI.

Adds gateway regression coverage.

* fix(tui): up-arrow inside a multi-line buffer moves cursor, not history

Reported during TUI v2 blitz retest: typing a multi-line message with
shift-Enter and then pressing Up to edit an earlier line swapped the
whole buffer for the previous history entry instead of moving the
cursor up a line.  Down then restored the draft → the buffer appeared
to "flip" between the draft and a prior prompt.

`useInputHandlers` cycles history on Up/Down, but textInput only
checked `inputBuf.length` — that only counts lines committed with a
trailing backslash, not shift-Enter newlines inside `input` itself.

Fix: detect logical lines inside the input string and move the cursor
one line up/down preserving column offset (clamp to line end when the
destination is shorter, standard editor behavior).  Only fall through
to history cycling when the cursor is already on the first line (Up)
or last line (Down).

Adds unit coverage for the new `lineNav` helper.

* fix(tui): /history shows the TUI's own transcript, scrollable

Reported during TUI v2 blitz retest: `/history` in the TUI only shows
prompts from non-TUI Hermes runs and can't scroll the window.  Root
cause is the slash-worker subprocess: it's a detached HermesCLI that
never sees the TUI's turns, so its `conversation_history` starts empty
and `show_history` surfaces whatever was persisted from earlier CLI
sessions — not what the user just did inside the TUI.

Intercept `/history` as a local slash command so it dumps
`ctx.local.getHistoryItems()` — the TUI's own transcript — routed
through the pager (which scrolls after #13591).  Accepts an optional
preview-length argument (default 400 chars per message).

Adds createSlashHandler coverage.

* fix(tui): tool inline_diff renders inline with the active turn

Reported during TUI v2 blitz retest: code-review diffs from tool.complete
appeared at the top of the current interaction thread, out of sequence
with the agent's messages and tool rows below them.

Root cause — `sys(inline_diff)` appends to `historyItems`, which sits
above the `StreamingAssistant` pane that renders the active turn.
Until the turn closed, the diff visually floated above everything
else happening in the same turn.

Route the diff through `turnController.appendSegmentMessage` instead
so it flushes any pending streaming text first, then lands in the
segment stream beside assistant output and tool calls.  On
`message.complete` the segment list is committed to history in emit
order (diff → final text), matching what the gateway sent.

Adds a regression test that exercises tool.complete → message.complete
with an inline_diff payload and asserts both the streaming and final
placement.

* feat(delegate): cross-agent file state coordination for concurrent subagents (#13718)

* feat(models): hide OpenRouter models that don't advertise tool support

Port from Kilo-Org/kilocode#9068.

hermes-agent is tool-calling-first — every provider path assumes the
model can invoke tools. Models whose OpenRouter supported_parameters
doesn't include 'tools' (e.g. image-only or completion-only models)
cannot be driven by the agent loop and fail at the first tool call.

Filter them out of fetch_openrouter_models() so they never appear in
the model picker (`hermes model`, setup wizard, /model slash command).

Permissive when the field is missing — OpenRouter-compatible gateways
(Nous Portal, private mirrors, older snapshots) don't always populate
supported_parameters. Treat missing as 'unknown → allow' rather than
silently emptying the picker on those gateways. Only hide models
whose supported_parameters is an explicit list that omits tools.

Tests cover: tools present → kept, tools absent → dropped, field
missing → kept, malformed non-list → kept, non-dict item → kept,
empty list → dropped.

* feat(delegate): cross-agent file state coordination for concurrent subagents

Prevents mangled edits when concurrent subagents touch the same file
(same process, same filesystem — the mangle scenario from #11215).

Three layers, all opt-out via HERMES_DISABLE_FILE_STATE_GUARD=1:

1. FileStateRegistry (tools/file_state.py) — process-wide singleton
   tracking per-agent read stamps and the last writer globally.
   check_stale() names the sibling subagent in the warning when a
   non-owning agent wrote after this agent's last read.

2. Per-path threading.Lock wrapped around the read-modify-write
   region in write_file_tool and patch_tool. Concurrent siblings on
   the same path serialize; different paths stay fully parallel.
   V4A multi-file patches lock in sorted path order (deadlock-free).

3. Delegate-completion reminder in tools/delegate_tool.py: after a
   subagent returns, writes_since(parent, child_start, parent_reads)
   appends '[NOTE: subagent modified files the parent previously
   read — re-read before editing: ...]' to entry.summary when the
   child touched anything the parent had already seen.

Complements (does not replace) the existing path-overlap check in
run_agent._should_parallelize_tool_batch — batch check prevents
same-file parallel dispatch within one agent's turn (cheap prevention,
zero API cost), registry catches cross-subagent and cross-turn
staleness at write time (detection).

Behavior is warning-only, not hard-failing — matches existing project
style. Errors surface naturally: sibling writes often invalidate the
old_string in patch operations, which already errors cleanly.

Tests: tests/tools/test_file_state_registry.py — 16 tests covering
registry state transitions, per-path locking, per-path-not-global
locking, writes_since filtering, kill switch, and end-to-end
integration through the real read_file/write_file/patch handlers.

* fix(tui): only cycle history at input boundaries on arrows

Follow-up on #13726 from blitz feedback: Up/Down history cycling should only trigger when the caret is at the start/end boundary (or the input is empty).\n\nPreviously useInputHandlers intercepted arrows whenever inputBuf was empty, which still stole Up/Down from normal multiline editing. textInput now publishes caret position through inputSelectionStore even with no active selection, and useInputHandlers gates history/queue cycling on those boundaries.

* fix(tui): keep inline diffs below tool rows and strip ANSI

Follow-up on #13729 from blitz screenshot feedback.\n\n- When tool.complete carried inline_diff but no buffered assistant text existed, pending tool rows were still in streamPendingTools, so diff rendered above the tool row section. appendSegmentMessage now emits pending tool rows as a trail segment before appending the diff artifact.\n- Strip ANSI color escapes from inline_diff payloads so we don't render loud red/green terminal palettes in the transcript.

* fix(tui): narrow /resume sources to human adapters

Follow-up on #13724: showing literally every source was too noisy.\n\n now fetches a wider window (, larger limit) and then filters to a curated allowlist of human-facing sources (tui/cli plus chat adapters like telegram/discord/slack/whatsapp/etc). This keeps row #7 fixed (telegram sessions visible in /resume) without surfacing internal source kinds such as tool/acp.

* fix(tui): arrow history fallback when no line exists

Follow-up on multiline arrow behavior: Up/Down now fall back to queue/history whenever there is no logical line above/below the caret (not only at absolute start/end character positions). This makes Up from the end of the top line cycle history, matching expected readline-ish behavior.

* fix(tui): render inline diffs inside assistant completion

Follow-up for #13729: segment-level system artifacts still looked detached in real flow.\n\nInstead of appending inline_diff as a standalone segment/system row, queue sanitized diffs during tool.complete and append them as a fenced diff block to the assistant completion text on message.complete. This keeps the diff in the same message flow as the assistant response.

* fix(tui): dedupe inline_diff when assistant already echoes it

Avoid duplicate diff rendering in #13729 flow. We now skip queued inline diffs that are already present in final assistant text and dedupe repeated queued diffs by exact content.

* fix(tui): keep review-diff tool rows terse

When tool.complete already carries inline_diff, the assistant message owns the full diff block. Suppress the tool-row summary/detail in that case so the turn shows one detailed diff surface instead of a rich diff plus a duplicated tool-detail payload.

* fix(tui): dedupe inline diffs, strip CLI review-diff header

After the prior inline-diff fix, the gateway still prepends a literal
"  ┊ review diff" line to inline_diff (it's terminal chrome written by
`_emit_inline_diff`). Wrapping that in a ```diff fence left that header
inside the code block. The agent also often narrates its own edit in a
second fenced diff, so the assistant message ended up stacking two
diff blocks for the same change.

- Strip the leading "┊ review diff" header from queued inline diffs
  before fencing.
- Skip appending the fenced diff entirely when the assistant already
  wrote its own ```diff (or ```patch) fence.

Keeps the single-surface diff UX even when the agent is chatty.

* fix(tts): use per-provider input-character caps instead of global 4000 (#13743)

A single global MAX_TEXT_LENGTH = 4000 truncated every TTS provider at
4000 chars, causing long inputs to be silently chopped even though the
underlying APIs allow much more:

  - OpenAI:     4096
  - xAI:        15000
  - MiniMax:    10000
  - ElevenLabs: 5000 / 10000 / 30000 / 40000 (model-aware)
  - Gemini:     ~5000
  - Edge:       ~5000

The schema description also told the model 'Keep under 4000 characters',
which encouraged the agent to self-chunk long briefs into multiple TTS
calls (producing 3 separate audio files instead of one).

New behavior:
  - PROVIDER_MAX_TEXT_LENGTH table + ELEVENLABS_MODEL_MAX_TEXT_LENGTH
    encode the documented per-provider limits.
  - _resolve_max_text_length(provider, cfg) resolves:
      1. tts.<provider>.max_text_length user override
      2. ElevenLabs model_id lookup
      3. provider default
      4. 4000 fallback
  - text_to_speech_tool() and stream_tts_to_speaker() both call the
    resolver; old MAX_TEXT_LENGTH alias kept for back-compat.
  - Schema description no longer hardcodes 4000.

Tests: 27 new unit + E2E tests; all 53 existing TTS tests and 253
voice-command/voice-cli tests still pass.

* feat(skills): add baoyu-comic skill

* refactor(skills): adapt baoyu-comic for Hermes

Port the upstream baoyu-comic skill to Hermes' tool ecosystem, matching
the earlier baoyu-infographic adaptation:

- metadata namespace openclaw -> hermes (+ tags, homepage)
- drop EXTEND.md preferences system (references/config/ removed,
  workflow Step 1.1 removed)
- user prompts via clarify (one question at a time) instead of
  AskUserQuestion batches
- image generation via image_generate instead of baoyu-imagine, with
  aspect-ratio mapping to landscape/portrait/square
- Windows/PowerShell/WSL shell snippets dropped
- file I/O referenced via Hermes write_file/read_file tools
- CLI-style --flags converted to natural-language options and
  user-intent cues (skill matching has no slash command trigger)

Add PORT_NOTES.md documenting the adaptations and a sync procedure.
Art-style/tone/layout reference files are preserved verbatim from
upstream v1.56.1.

* fix(skills): address baoyu-comic PR review

- Remove PDF merge feature and scripts/ directory (no pdf-lib dep)
- Correct image_generate docs: prompt-only, returns URL; add
  curl download step after every call
- Downgrade reference images to text-based trait extraction
  (style/palette/scene); character sheet is agent-facing reference
- Unify source file naming on source-{slug}.md across SKILL.md
  and workflow.md

* fix(skills): clarify baoyu-comic character sheet role

Page prompts are written in Step 5 from the text descriptions in
characters/characters.md — the PNG sheet generated in Step 7.1
cannot be used to write them. Reposition the PNG as a human-facing
review artifact (and reference for later regenerations / manual
edits), and drop the confusing "Character sheet | Strategy" tables
since the embedding rule is uniform.

* docs: document delegation width + depth knobs (#13745)

Fills the three gaps left by the orchestrator/width-depth salvage:

- configuration.md §Delegation: max_concurrent_children, max_spawn_depth,
  orchestrator_enabled are now in the canonical config.yaml reference
  with a paragraph covering defaults, clamping, role-degradation, and
  the 3x3x3=27-leaf cost scaling.
- environment-variables.md: adds DELEGATION_MAX_CONCURRENT_CHILDREN to
  the Agent Behavior table.
- features/delegation.md: corrects stale 'default 5, cap 8' wording
  (that was from the original PR; the salvage landed on default 3 with
  no ceiling and a tool error on excess instead of truncation).

* fix(website): run skill extraction automatically on npm run build/start (#13747)

website/src/pages/skills/index.tsx imports ../../data/skills.json, but
that file is git-ignored and generated at build time by
website/scripts/extract-skills.py. CI workflows (deploy-site.yml,
docs-site-checks.yml) run the script explicitly before 'npm run build',
so production and PR checks always work — but 'npm run build' on a
contributor's machine fails with:

  Module not found: Can't resolve '../../data/skills.json'

because the extraction step was never wired into the npm scripts.

Adds a prebuild/prestart hook that runs extract-skills.py automatically.
If python3 or pyyaml aren't installed locally, writes an empty
skills.json instead of hard-failing — the Skills Hub page renders with
an empty state, the rest of the site builds normally, and CI (which
always has the deps) still generates the full catalog for production.

* fix(skills/baoyu-comic): absolute curl paths + clarify-timeout handling (#13775)

* fix(skills/baoyu-comic): require absolute paths for curl -o downloads

When downloading generated images across several batches of image_generate
calls, relying on persistent-shell CWD is unsafe. The terminal tool's shell
can rotate (TERMINAL_LIFETIME_SECONDS expiry, a failed cd that leaves the
shell somewhere else), and 'curl -fsSL <url> -o relative.png' then silently
writes to the wrong directory with no error.

Update the skill's Step 7 Download step to require absolute -o paths (or
workdir= on the terminal tool) and add a matching pitfall entry referencing
the Apr 2026 incident where pages 06-09 of a 10-page comic landed at the
repo root instead of comic/<slug>/. The agent then spent several turns
claiming the files existed where they didn't.

* fix(skills/baoyu-comic): handle clarify timeouts correctly in Step 2

A clarify timeout returning 'Use your best judgement to make the choice
and proceed' is NOT user consent to default the entire Step 2 questionnaire.
It is a per-question default only. Add guidance at both instruction sites
(SKILL.md User Questions section, references/workflow.md Step 2 header)
telling the agent to:

1. Continue asking the remaining questions in the sequence after a
   timeout — each question is an independent consent point.
2. Surface every defaulted choice in the next user-visible message
   so the user can correct it when they return. An unreported default
   is indistinguishable from never having asked.

Reported live Apr 2026: agent asked style question via clarify, got a
timeout response, and silently defaulted style + narrative focus +
audience + review flags in one pass. User only learned style had
defaulted to 'ohmsha' after the comic was fully generated.

* fix(prompt): tell CLI agents not to emit MEDIA:/path tags (#13766)

The CLI has no attachment channel — MEDIA:<path> tags are only
intercepted on messaging gateway platforms (Telegram, Discord,
Slack, WhatsApp, Signal, BlueBubbles, email, etc.). On the CLI
they render as literal text, which is confusing for users.

The CLI platform hint was the one PLATFORM_HINTS entry that said
nothing about file delivery, so models trained on the messaging
hints would default to MEDIA: tags on the CLI too. Tool schemas
(browser_tool, tts_tool, etc.) also recommend MEDIA: generically.

Extend the CLI hint to explicitly discourage MEDIA: tags and tell
the agent to reference files by plain absolute path instead.

Add a regression test asserting the CLI hint carries negative
guidance about MEDIA: while messaging hints keep positive guidance.

* fix: add User-Agent claude-code/0.1.0 for Kimi /coding endpoint

- Add _is_kimi_coding_endpoint() to detect Kimi coding API
- Place Kimi check BEFORE _requires_bearer_auth to ensure User-Agent header is set
- Without this header, Kimi returns 403 on /coding/v1/messages
- Fixes kimi-2.5, kimi-for-coding, kimi-k2.6-code-preview all returning 403

* fix: auto-detect anthropic_messages mode for Kimi /coding/v1 endpoints

* fix(kimi-coding): add KIMI_CODING_API_KEY fallback + api_mode detection for /coding endpoint

* fix(kimi-coding): set anthropic_messages api_mode for /coding endpoint

* fix: Update Kimi Coding API endpoint and User-Agent

* fix: Enhance Kimi Coding API mode detection and User-Agent

* fix(kimi): reconcile sk-kimi- routing with Anthropic SDK URL semantics

Follow-ups after salvaging xiaoqiang243's kimi-for-coding patches:

- KIMI_CODE_BASE_URL: drop trailing /v1 (was /coding/v1).
  The /coding endpoint speaks Anthropic Messages, and the Anthropic SDK
  appends /v1/messages internally. /coding/v1 + SDK suffix produced
  /coding/v1/v1/messages (a 404). /coding + SDK suffix now yields
  /coding/v1/messages correctly.
- kimi-coding ProviderConfig: keep legacy default api.moonshot.ai/v1 so
  non-sk-kimi- moonshot keys still authenticate. sk-kimi- keys are
  already redirected to api.kimi.com/coding via _resolve_kimi_base_url.
- doctor.py: update Kimi UA to claude-code/0.1.0 (was KimiCLI/1.30.0)
  and rewrite /coding base URLs to /coding/v1 for the /models health
  check (Anthropic surface has no /models).
- test_kimi_env_vars: accept KIMI_CODING_API_KEY as a secondary env var.

E2E verified:
  sk-kimi-<key>  → https://api.kimi.com/coding/v1/messages (Anthropic)
  sk-<legacy>    → https://api.moonshot.ai/v1/chat/completions (OpenAI)
  UA: claude-code/0.1.0, x-api-key: <sk-kimi-*>

* chore(release): map xiaoqiang243 personal email in AUTHOR_MAP

* feat: add ResponsesApiTransport + wire all Codex transport paths

Add ResponsesApiTransport wrapping codex_responses_adapter.py behind the
ProviderTransport ABC. Auto-registered via _discover_transports().

Wire ALL Codex transport methods to production paths in run_agent.py:
- build_kwargs: main _build_api_kwargs codex branch (50 lines extracted)
- normalize_response: main loop + flush + summary + retry (4 sites)
- convert_tools: memory flush tool override
- convert_messages: called internally via build_kwargs
- validate_response: response validation gate
- preflight_kwargs: request sanitization (2 sites)

Remove 7 dead legacy wrappers from AIAgent (_responses_tools,
_chat_messages_to_responses_input, _normalize_codex_response,
_preflight_codex_api_kwargs, _preflight_codex_input_items,
_extract_responses_message_text, _extract_responses_reasoning_text).
Keep 3 ID manipulation methods still used by _build_assistant_message.

Update 18 test call sites across 3 test files to call adapter functions
directly instead of through deleted AIAgent wrappers.

24 new tests. 343 codex/responses/transport tests pass (0 failures).

PR 4 of the provider transport refactor.

* fix(delegation): add hard timeout and stale detection for subagent execution (#13770)

- Wrap child.run_conversation() in a ThreadPoolExecutor with configurable
  timeout (delegation.child_timeout_seconds, default 300s) to prevent
  indefinite blocking when a subagent's API call or tool HTTP request hangs.

- Add heartbeat stale detection: if a child's api_call_count doesn't
  advance for 5 consecutive heartbeat cycles (~2.5 min), stop touching
  the parent's activity timestamp so the gateway inactivity timeout
  can fire as a last resort.

- Add 'timeout' as a new exit_reason/status alongside the existing
  completed/max_iterations/interrupted states.

- Use shutdown(wait=False) on the timeout executor to avoid the
  ThreadPoolExecutor.__exit__ deadlock when a child is stuck on
  blocking I/O.

Closes #13768

* remove Nous Portal free-model allowlist

Drop _NOUS_ALLOWED_FREE_MODELS + filter_nous_free_models and its two call
sites. Whatever Nous Portal prices as free now shows up in the picker as-is
— no local allowlist gatekeeping. Free-tier partitioning (paid vs free in
the menu) still runs via partition_nous_models_by_tier.

* feat(aux): use Portal /api/nous/recommended-models for auxiliary models

Wire the auxiliary client (compaction, vision, session search, web extract)
to the Nous Portal's curated recommended-models endpoint when running on
Nous Portal, with a TTL-cached fetch that mirrors how we pull /models for
pricing.

hermes_cli/models.py
  - fetch_nous_recommended_models(portal_base_url, force_refresh=False)
    10-minute TTL cache, keyed per portal URL (staging vs prod don't
    collide).  Public endpoint, no auth required.  Returns {} on any
    failure so callers always get a dict.
  - get_nous_recommended_aux_model(vision, free_tier=None, ...)
    Tier-aware pick from the payload:
      - Paid tier → paidRecommended{Vision,Compaction}Model, falling back
        to freeRecommended* when the paid field is null (common during
        staged rollouts of new paid models).
      - Free tier → freeRecommended* only, never leaks paid models.
    When free_tier is None, auto-detects via the existing
    check_nous_free_tier() helper (already cached 3 min against
    /api/oauth/account).  Detection errors default to paid so we never
    silently downgrade a paying user.

agent/auxiliary_client.py — _try_nous()
  - Replaces the hardcoded xiaomi/mimo free-tier branch with a single call
    to get_nous_recommended_aux_model(vision=vision).
  - Falls back to _NOUS_MODEL (google/gemini-3-flash-preview) when the
    Portal is unreachable or returns a null recommendation.
  - The Portal is now the source of truth for aux model selection; the
    xiaomi allowlist we used to carry is effectively dead.

Tests (15 new)
  - tests/hermes_cli/test_models.py::TestNousRecommendedModels
    Fetch caching, per-portal keying, network failure, force_refresh;
    paid-prefers-paid, paid-falls-to-free, free-never-leaks-paid,
    auto-detect, detection-error → paid default, null/blank modelName
    handling.
  - tests/agent/test_auxiliary_client.py::TestNousAuxiliaryRefresh
    _try_nous honors Portal recommendation for text + vision, falls
    back to google/gemini-3-flash-preview on None or exception.

Behavior won't visibly change today — both tier recommendations currently
point at google/gemini-3-flash-preview — but the moment the Portal ships
a better paid recommendation, subscribers pick it up within 10 minutes
without a Hermes release.

* feat: add ChatCompletionsTransport + wire all default paths

Third concrete transport — handles the default 'chat_completions' api_mode used
by ~16 OpenAI-compatible providers (OpenRouter, Nous, NVIDIA, Qwen, Ollama,
DeepSeek, xAI, Kimi, custom, etc.). Wires build_kwargs + validate_response to
production paths.

Based on PR #13447 by @kshitijk4poor, with fixes:
- Preserve tool_call.extra_content (Gemini thought_signature) via
  ToolCall.provider_data — the original shim stripped it, causing 400 errors
  on multi-turn Gemini 3 thinking requests.
- Preserve reasoning_content distinctly from reasoning (DeepSeek/Moonshot) so
  the thinking-prefill retry check (_has_structured) still triggers.
- Port Kimi/Moonshot quirks (32000 max_tokens, top-level reasoning_effort,
  extra_body.thinking) that landed on main after the original PR was opened.
- Keep _qwen_prepare_chat_messages_inplace alive and call it through the
  transport when sanitization already deepcopied (avoids a second deepcopy).
- Skip the back-compat SimpleNamespace shim in the main normalize loop — for
  chat_completions, response.choices[0].message is already the right shape
  with .content/.tool_calls/.reasoning/.reasoning_content/.reasoning_details
  and per-tool-call .extra_content from the OpenAI SDK.

run_agent.py: -239 lines in _build_api_kwargs default branch extracted to the
transport. build_kwargs now owns: codex-field sanitization, Qwen portal prep,
developer role swap, provider preferences, max_tokens resolution (ephemeral >
user > NVIDIA 16384 > Qwen 65536 > Kimi 32000 > anthropic_max_output), Kimi
reasoning_effort + extra_body.thinking, OpenRouter/Nous/GitHub reasoning,
Nous product attribution tags, Ollama num_ctx, custom-provider think=false,
Qwen vl_high_resolution_images, request_overrides.

39 new transport tests (8 build_kwargs, 5 Kimi, 4 validate, 4 normalize
including extra_content regression, 3 cache stats, 3 basic). Tests/run_agent/
targeted suite passes (885/885 + 15 skipped; the 1 remaining failure is the
test_concurrent_interrupt flake present on origin/main).

* fix(tui): don't force-open Activity on every error

Reverts the auto-expand-on-new-error effect added in 93b47d96. The
effect overrode the user's chosen detailsMode and visually interrupted
every turn. Red/yellow chevron tint remains as the passive signal —
click to read, just like Thinking and Tool calls.

* fix(tui): demote gateway log-noise from Activity to info tone

Restore the old-CLI contract where only complete failures tint Activity
red. Everything else is still visible for debugging but no longer
commandeers attention.

- gateway.stderr: always tone='info' (drops the ERRLIKE_RE regex)
- gateway.protocol_error: both pushes demoted to 'info'
- commands.catalog cold-start failure: demoted to 'info'
- approval.request: no longer duplicates the overlay into Activity

Kept as 'error': terminal `error` event, gateway.start_timeout,
gateway-exited, explicit status.update kinds.

* feat: add BedrockTransport + wire all Bedrock transport paths

Fourth and final transport — completes the transport layer with all four
api_modes covered.  Wraps agent/bedrock_adapter.py behind the ProviderTransport
ABC, handles both raw boto3 dicts and already-normalized SimpleNamespace.

Wires all transport methods to production paths in run_agent.py:
- build_kwargs: _build_api_kwargs bedrock branch
- validate_response: response validation, new bedrock_converse branch
- finish_reason: new bedrock_converse branch in finish_reason extraction

Based on PR #13467 by @kshitijk4poor, with one adjustment: the main normalize
loop does NOT add a bedrock_converse branch to invoke normalize_response on
the already-normalized response.  Bedrock's normalize_converse_response runs
at the dispatch site (run_agent.py:5189), so the response already has the
OpenAI-compatible .choices[0].message shape by the time the main loop sees
it.  Falling through to the chat_completions else branch is correct and
sidesteps a redundant NormalizedResponse rebuild.

Transport coverage — complete:
| api_mode           | Transport                | build_kwargs | normalize | validate |
|--------------------|--------------------------|:------------:|:---------:|:--------:|
| anthropic_messages | AnthropicTransport       | ✅            | ✅         | ✅        |
| codex_responses    | ResponsesApiTransport    | ✅            | ✅         | ✅        |
| chat_completions   | ChatCompletionsTransport | ✅            | ✅         | ✅        |
| bedrock_converse   | BedrockTransport         | ✅            | ✅         | ✅        |

17 new BedrockTransport tests pass.  117 transport tests total pass.
160 bedrock/converse tests across tests/agent/ pass.  Full tests/run_agent/
targeted suite passes (885/885 + 15 skipped; the 1 remaining failure is the
pre-existing test_concurrent_interrupt flake on origin/main).

* chore(models): drop 3 models from nous portal recommended list (#13822)

Remove nvidia/nemotron-3-super-120b-a12b:free, arcee-ai/trinity-large-preview:free,
and openrouter/elephant-alpha from _PROVIDER_MODELS['nous']. The paid nemotron and
arcee-thinking variants remain.

* fix(kimi): don't send Anthropic thinking to api.kimi.com/coding (#13826)

Kimi's /coding endpoint speaks the Anthropic Messages protocol but has
its own thinking semantics: when thinking.enabled is sent, Kimi validates
the history and requires every prior assistant tool-call message to carry
OpenAI-style reasoning_content. The Anthropic path never populates that
field, and convert_messages_to_anthropic strips Anthropic thinking blocks
on third-party endpoints — so after one tool-calling turn the next request
fails with:

  HTTP 400: thinking is enabled but reasoning_content is missing in
  assistant tool call message at index N

Kimi on chat_completions handles thinking via extra_body in
ChatCompletionsTransport (#13503). On the Anthropic route, drop the
parameter entirely and let Kimi drive reasoning server-side.

build_anthropic_kwargs now gates the reasoning_config -> thinking block
on not _is_kimi_coding_endpoint(base_url).

Tests: 8 new parametric tests cover /coding, /coding/v1, /coding/anthropic,
/coding/ (trailing slash), explicit disabled, other third-party endpoints
still getting thinking (MiniMax), native Anthropic unaffected, and the
non-/coding Kimi root route.

* feat(models): add minimax/minimax-m2.5:free to OpenRouter catalog (#13836)

Surfaces the free variant alongside the paid minimax-m2.5 entry in
both the OPENROUTER_MODELS fallback snapshot and the nous/openrouter
provider model list.

* feat(plugins): pluggable image_gen backends + OpenAI provider (#13799)

* feat(plugins): pluggable image_gen backends + OpenAI provider

Adds a ImageGenProvider ABC so image generation backends register as
bundled plugins under `plugins/image_gen/<name>/`. The plugin scanner
gains three primitives to make this work generically:

- `kind:` manifest field (`standalone` | `backend` | `exclusive`).
  Bundled `kind: backend` plugins auto-load — no `plugins.enabled`
  incantation. User-installed backends stay opt-in.
- Path-derived keys: `plugins/image_gen/openai/` gets key
  `image_gen/openai`, so a future `tts/openai` cannot collide.
- Depth-2 recursion into category namespaces (parent dirs without a
  `plugin.yaml` of their own).

Includes `OpenAIImageGenProvider` as the first consumer (gpt-image-1.5
default, plus gpt-image-1, gpt-image-1-mini, DALL-E 3/2). Base64
responses save to `$HERMES_HOME/cache/images/`; URL responses pass
through.

FAL stays in-tree for this PR — a follow-up ports it into
`plugins/image_gen/fal/` so the in-tree `image_generation_tool.py`
slims down. The dispatch shim in `_handle_image_generate` only fires
when `image_gen.provider` is explicitly set to a non-FAL value, so
existing FAL setups are untouched.

- 41 unit tests (scanner recursion, kind parsing, gate logic,
  registry, OpenAI payload shapes)
- E2E smoke verified: bundled plugin autoloads, registers, and
  `_handle_image_generate` routes to OpenAI when configured

* fix(image_gen/openai): don't send response_format to gpt-image-*

The live API rejects it: 'Unknown parameter: response_format'
(verified 2026-04-21 with gpt-image-1.5). gpt-image-* models return
b64_json unconditionally, so the parameter was both unnecessary and
actively broken.

* feat(image_gen/openai): gpt-image-2 only, drop legacy catalog

gpt-image-2 is the latest/best OpenAI image model (released 2026-04-21)
and there's no reason to expose the older gpt-image-1.5 / gpt-image-1 /
dall-e-3 / dall-e-2 alongside it — slower, lower quality, or awkward
(dall-e-2 squares only). Trim the catalog down to a single model.

Live-verified end-to-end: landscape 1536x1024 render of a Moog-style
synth matches prompt exactly, 2.4MB PNG saved to cache.

* feat(image_gen/openai): expose gpt-image-2 as three quality tiers

Users pick speed/fidelity via the normal model picker instead of a
hidden quality knob. All three tier IDs resolve to the single underlying
gpt-image-2 API model with a different quality parameter:

  gpt-image-2-low     ~15s   fast iteration
  gpt-image-2-medium  ~40s   default
  gpt-image-2-high    ~2min  highest fidelity

Live-measured on OpenAI's API today: 15.4s / 40.8s / 116.9s for the
same 1024x1024 prompt.

Config:
  image_gen.openai.model: gpt-image-2-high
  # or
  image_gen.model: gpt-image-2-low
  # or env var for scripts/tests
  OPENAI_IMAGE_MODEL=gpt-image-2-medium

Live-verified end-to-end with the low tier: 18.8s landscape render of a
golden retriever in wildflowers, vision-confirmed exact match.

* feat(tools_config): plugin image_gen providers inject themselves into picker

'hermes tools' → Image Generation now shows plugin-registered backends
alongside Nous Subscription and FAL.ai without tools_config.py needing
to know about them. OpenAI appears as a third option today; future
backends appear automatically as they're added.

Mechanism:
- ImageGenProvider gains an optional get_setup_schema() hook
  (name, badge, tag, env_vars). Default derived from display_name.
- tools_config._plugin_image_gen_providers() pulls the schemas from
  every registered non-FAL plugin provider.
- _visible_providers() appends those rows when rendering the Image
  Generation category.
- _configure_provider() handles the new image_gen_plugin_name marker:
  writes image_gen.provider and routes to the plugin's list_models()
  catalog for the model picker.
- _toolset_needs_configuration_prompt('image_gen') stops demanding a
  FAL key when any plugin provider reports is_available().

FAL is skipped in the plugin path because it already has hardcoded
TOOL_CATEGORIES rows — when it gets ported to a plugin in a follow-up
PR the hardcoded rows go away and it surfaces through the same path
as OpenAI.

Verified live: picker shows Nous Subscription / FAL.ai / OpenAI.
Picking OpenAI prompts for OPENAI_API_KEY, then shows the
gpt-image-2-low/medium/high model picker sourced from the plugin.

397 tests pass across plugins/, tools_config, registry, and picker.

* fix(image_gen): close final gaps for plugin-backend parity with FAL

Two small places that still hardcoded FAL:

- hermes_cli/setup.py status line: an OpenAI-only setup showed
  'Image Generation: missing FAL_KEY'. Now probes plugin providers
  and reports '(OpenAI)' when one is_available() — or falls back to
  'missing FAL_KEY or OPENAI_API_KEY' if nothing is configured.

- image_generate tool schema description: said 'using FAL.ai, default
  FLUX 2 Klein 9B'. Rewrote provider-neutral — 'backend and model are
  user-configured' — and notes the 'image' field can be a URL or an
  absolute path, which the gateway delivers either way via
  extract_local_files().

* feat: add Step Plan provider support (salvage #6005)

Adds a first-class 'stepfun' API-key provider surfaced as Step Plan:

- Support Step Plan setup for both International and China regions
- Discover Step Plan models live from /step_plan/v1/models, with a
  small coding-focused fallback catalog when discovery is unavailable
- Thread StepFun through provider metadata, setup persistence, status
  and doctor output, auxiliary routing, and model normalization
- Add tests for provider resolution, model validation, metadata
  mapping, and StepFun region/model persistence

Based on #6005 by @hengm3467.

Co-authored-by: hengm3467 <100685635+hengm3467@users.noreply.github.com>

* fix(packaging): include agent.* sub-packages in pyproject.toml

The transport refactor (PRs #13862 ff.) added agent/transports/ as a
sub-package but the setuptools packages.find include list only had
"agent" (top-level files), not "agent.*" (sub-packages).

pip install / Nix builds therefore ship run_agent.py (which now imports
from agent.transports on every API call) but omit the transports
directory entirely, causing:

  ModuleNotFoundError: No module named 'agent.transports'

on every LLM call for packaged installs.

Adds "agent.*" to match the existing pattern used by tools, gateway,
tui_gateway, and plugins.

* fix: preserve reasoning_content on Kimi replay

* feat(optional-skills): add page-agent skill under new web-development category (#13976)

Adds an optional skill that walks users through installing and using
alibaba/page-agent — a pure-JS in-page GUI agent that web developers
embed into their own webapps so end users can drive the UI with
natural language.

Three install paths: CDN demo (30s, no install), npm install into an
existing app with provider config table (Qwen/OpenAI/Ollama/OpenRouter),
and clone-from-source for dev/contributor workflow.

Clear use-case framing up front (embed AI copilot in SaaS/admin/B2B,
modernize legacy UIs, accessibility via natural language) and an
explicit NOT-for list that points users wanting server-side browser
automation back to Hermes' built-in browser tool.

Live-verified: repo builds on Node 22.22 + npm 10.9, dev:demo serves
at localhost:5174, API surface (new PageAgent{...}, panel.show(),
execute(task)) matches what the skill documents. Also verified
discovery end-to-end via OptionalSkillSource with isolated
HERMES_HOME — search/inspect/fetch all resolve
official/web-development/page-agent correctly.

New category directory: optional-skills/web-development/ with a
DESCRIPTION.md explaining the distinction from Hermes' own browser
automation (outside-in vs inside-out).

* feat(wecom): add QR scan flow and interactive setup wizard for bot credentials

* docs(wecom): document QR scan-to-create setup flow

* fix(wecom): visible poll progress + clearer no-bot-info failure + docstring note

Follow-ups on top of salvaged #13923 (@keifergu):
- Print QR poll dot every 3s instead of every 18s so "Fetching
  configuration results..." doesn't look hung.
- On "status=success but no bot_info" from the WeCom query endpoint,
  log the full payload at WARNING and tell the user we're falling
  back to manual entry (was previously a single opaque line).
- Document in the qr_scan_for_bot_info() docstring that the
  work.weixin.qq.com/ai/qc/* endpoints are the admin-console web-UI
  flow, not the public developer API, and may change without notice.

Also add keifergu@tencent.com to scripts/release.py AUTHOR_MAP so
release notes attribute the feature correctly.

* feat(state): auto-prune old sessions + VACUUM state.db at startup (#13861)

* feat(state): auto-prune old sessions + VACUUM state.db at startup

state.db accumulates every session, message, and FTS5 index entry forever.
A heavy user (gateway + cron) reported 384MB with 982 sessions / 68K messages
causing slowdown; manual 'hermes sessions prune --older-than 7' + VACUUM
brought it to 43MB. The prune command and VACUUM are not wired to run
automatically anywhere — sessions grew unbounded until users noticed.

Changes:
- hermes_state.py: new state_meta key/value table, vacuum() method, and
  maybe_auto_prune_and_vacuum() — idempotent via last-run timestamp in
  state_meta so it only actually executes once per min_interval_hours
  across all Hermes processes for a given HERMES_HOME. Never raises.
- hermes_cli/config.py: new 'sessions:' block in DEFAULT_CONFIG
  (auto_prune=True, retention_days=90, vacuum_after_prune=True,
  min_interval_hours=24). Added to _KNOWN_ROOT_KEYS.
- cli.py: call maintenance once at HermesCLI init (shared helper
  _run_state_db_auto_maintenance reads config and delegates to DB).
- gateway/run.py: call maintenance once at GatewayRunner init.
- Docs: user-guide/sessions.md rewrites 'Automatic Cleanup' section.

Why VACUUM matters: SQLite does NOT shrink the file on DELETE — freed
pages get reused on next INSERT. Without VACUUM, a delete-heavy DB stays
bloated forever. VACUUM only runs when the prune actually removed rows,
so tight DBs don't pay the I/O cost.

Tests: 10 new tests in tests/test_hermes_state.py covering state_meta,
vacuum, idempotency, interval skipping, VACUUM-only-when-needed,
corrupt-marker recovery. All 246 existing state/config/gateway tests
still pass.

Verified E2E with real imports + isolated HERMES_HOME: DEFAULT_CONFIG
exposes the new block, load_config() returns it for fresh installs,
first call prunes+vacuums, second call within min_interval_hours skips,
and the state_meta marker persists across connection close/reopen.

* sessions.auto_prune defaults to false (opt-in)

Session history powers session_search recall across past conversations,
so silently pruning on startup could surprise users. Ship the machinery
disabled and let users opt in when they notice state.db is hurting
performance.

- DEFAULT_CONFIG.sessions.auto_prune: True → False
- Call-site fallbacks in cli.py and gateway/run.py match the new default
  (so unmigrated configs still see off)
- Docs: flip 'Enable in config.yaml' framing + tip explains the tradeoff

* feat(hindsight): richer session-scoped retain metadata

- Add configurable retain_tags / retain_source / retain_user_prefix /
  retain_assistant_prefix knobs for native Hindsight.
- Thread gateway session identity (user_name, chat_id, chat_name,
  chat_type, thread_id) through AIAgent and MemoryManager into
  MemoryProvider.initialize kwargs so providers can scope and tag
  retained memories.
- Hindsight attaches the new identity fields as retain metadata,
  merges per-call tool tags with configured default tags, and uses
  the configurable transcript labels for auto-retained turns.

Co-authored-by: Abner <abner.the.foreman@agentmail.to>

* chore(release): map Abner email to Abnertheforeman

* refactor(qqbot): migrate qr onboard flow to sync + consolidate into onboard.py

- Replace async create_bind_task/poll_bind_result with synchronous
  httpx.Client equivalents, eliminating manual event loop management
- Move _render_qr and full qr_register() entry-point into onboard.py,
  mirroring the Feishu onboarding pattern
- Remove _qqbot_render_qr and _qqbot_qr_flow from gateway.py (~90 lines);
  call site becomes a single qr_register() import
- Fix potential segfault: previous code called loop.close() in the EXPIRED
  branch and again in the finally block (double-close crashed under uvloop)

* fix(cli): ensure project .env is sanitized before loading

* chore(release): map hharry11 email to GitHub handle

* feat(dashboard): track real API call count per session

Adds schema v7 'api_call_count' column. run_agent.py increments it by 1
per LLM API call, web_server analytics SQL aggregates it, frontend uses
the real counter instead of summing sessions.

The 'API Calls' card on the analytics dashboard previously displayed
COUNT(*) from the sessions table — the number of conversations, not
LLM requests. Each session makes 10-90 API calls through the tool loop,
so the reported number was ~30x lower than real.

Salvaged from PR #10140 (@kshitijk4poor). The cache-token accuracy
portions of the original PR were deferred — per-provider analytics is
the better path there, since cache_write_tokens and actual_cost_usd
are only reliably available from a subset of providers (Anthropic
native, Codex Responses, OpenRouter with usage.include).

Tests:
- schema_version v7 assertion
- migration v2 -> v7 adds api_call_count column with default 0
- update_token_counts increments api_call_count by provided delta
- absolute=True sets api_call_count directly
- /api/analytics/usage exposes total_api_calls in totals

* fix(plugins+nous): auto-coerce memory plugins; actionable Nous 401 diagnostic (#14005)

* fix(plugins): auto-coerce user-installed memory plugins to kind=exclusive

User-installed memory provider plugins…

NousResearch#13743)

A single global MAX_TEXT_LENGTH = 4000 truncated every TTS provider at
4000 chars, causing long inputs to be silently chopped even though the
underlying APIs allow much more:

  - OpenAI:     4096
  - xAI:        15000
  - MiniMax:    10000
  - ElevenLabs: 5000 / 10000 / 30000 / 40000 (model-aware)
  - Gemini:     ~5000
  - Edge:       ~5000

The schema description also told the model 'Keep under 4000 characters',
which encouraged the agent to self-chunk long briefs into multiple TTS
calls (producing 3 separate audio files instead of one).

New behavior:
  - PROVIDER_MAX_TEXT_LENGTH table + ELEVENLABS_MODEL_MAX_TEXT_LENGTH
    encode the documented per-provider limits.
  - _resolve_max_text_length(provider, cfg) resolves:
      1. tts.<provider>.max_text_length user override
      2. ElevenLabs model_id lookup
      3. provider default
      4. 4000 fallback
  - text_to_speech_tool() and stream_tts_to_speaker() both call the
    resolver; old MAX_TEXT_LENGTH alias kept for back-compat.
  - Schema description no longer hardcodes 4000.

Tests: 27 new unit + E2E tests; all 53 existing TTS tests and 253
voice-command/voice-cli tests still pass.

NousResearch#13743)

A single global MAX_TEXT_LENGTH = 4000 truncated every TTS provider at
4000 chars, causing long inputs to be silently chopped even though the
underlying APIs allow much more:

  - OpenAI:     4096
  - xAI:        15000
  - MiniMax:    10000
  - ElevenLabs: 5000 / 10000 / 30000 / 40000 (model-aware)
  - Gemini:     ~5000
  - Edge:       ~5000

The schema description also told the model 'Keep under 4000 characters',
which encouraged the agent to self-chunk long briefs into multiple TTS
calls (producing 3 separate audio files instead of one).

New behavior:
  - PROVIDER_MAX_TEXT_LENGTH table + ELEVENLABS_MODEL_MAX_TEXT_LENGTH
    encode the documented per-provider limits.
  - _resolve_max_text_length(provider, cfg) resolves:
      1. tts.<provider>.max_text_length user override
      2. ElevenLabs model_id lookup
      3. provider default
      4. 4000 fallback
  - text_to_speech_tool() and stream_tts_to_speaker() both call the
    resolver; old MAX_TEXT_LENGTH alias kept for back-compat.
  - Schema description no longer hardcodes 4000.

Tests: 27 new unit + E2E tests; all 53 existing TTS tests and 253
voice-command/voice-cli tests still pass.

PR #13743 replaced the global MAX_TEXT_LENGTH=4000 with a per-provider
table and a user-override 'max_text_length:' key, but the user-guide
TTS page documented no length behaviour at all. Users hitting truncation
had no way to discover the new caps or the override.

Add an 'Input length limits' subsection after the existing Configuration
YAML block: provider default caps (Edge 5000 / OpenAI 4096 / xAI 15000 /
MiniMax 10000 / Mistral 4000 / Gemini 5000 / ElevenLabs model-aware /
NeuTTS,KittenTTS 2000), ElevenLabs model_id -> cap table (5k-40k), an
override example, and the validation rules (non-positive / non-integer /
boolean values fall through to the provider default).

* fix(aux): trigger fallback on 429 rate-limit errors in auxiliary client

When a provider returns a 429 rate-limit error (not billing-related),
the auxiliary client's call_llm/async_call_llm previously did NOT trigger
the fallback chain. This caused auxiliary tasks like session_search to
exhaust all 3 retries against the same rate-limited endpoint, losing
session metadata that depended on the summarization completing.

Root cause: `_is_payment_error()` only matched 429s containing billing
keywords ("credits", "insufficient funds", etc.). Provider-specific
rate-limit messages like Nous's "Hold up for a bit, you've exceeded the
rate limit on your API key" didn't match, so `_is_payment_error` returned
False, `_is_connection_error` returned False, and `should_fallback` was
False — all retries hit the same rate-limited provider.

Fix:
- New `_is_rate_limit_error()` function that detects 429 + rate-limit
  keywords, generic 429 without billing keywords, and OpenAI SDK
  `RateLimitError` class instances (which may omit .status_code).
- Updated `should_fallback` in both `call_llm` and `async_call_llm` to
  include `_is_rate_limit_error`.
- Updated the max_tokens retry path to also check for rate-limit errors.
- Updated the reason string to include "rate limit".

This complements the Nous rate guard (PR #10568) which prevents new calls
to Nous when already rate-limited — this fix handles the case where a
request is already in flight when the 429 arrives.

Related: #8023, #12554, #11034
Co-authored-by: Zeejay <zjtan1@gmail.com>

* chore: AUTHOR_MAP entry for zeejaytan

* fix(acp): preserve assistant reasoning metadata in session persistence

* chore: AUTHOR_MAP entry for Aslaaen

* feat(cli): add list_picker_providers for credential-filtered picker

The Telegram/Discord /model pickers currently call
list_authenticated_providers(), which returns every provider whose
credentials resolve locally and every model in its curated snapshot.
Two failure modes fall out:

- OpenRouter rows can include IDs the live catalog no longer carries.
- Provider rows can surface with zero callable models (e.g. a slug
  whose credential pool entry exists but has nothing behind it).

list_picker_providers() wraps the base function and post-processes the
result so the interactive picker only shows models the user can
actually select:

- OpenRouter's models come from fetch_openrouter_models() (live-catalog
  filtered against the curated OPENROUTER_MODELS snapshot).
- Rows with an empty models list are dropped, except custom endpoints
  (is_user_defined=True with an api_url) where the user may enter
  model ids manually.
- All other fields pass through unchanged.

The gateway /model handler switches to the new helper for the
interactive picker payload only. Typed /model <name> and the text
fallback list stay on list_authenticated_providers() so nothing is
hidden from power users or platforms without a picker.

Covered by nine focused unit tests in
tests/hermes_cli/test_list_picker_providers.py.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

* chore: AUTHOR_MAP entry for Tkander1715

* feat(tui): remove /provider alias for /model (#20358)

/model is the canonical command; /provider was a redundant alias that
dispatched to the same ModelPicker overlay. Drop the alias, the regex
branch in useCompletion, and the alias-coverage test.

* fix: resolve lazy session creation regressions (#18370 fallout) (#20363)

Fix three regressions introduced by PR #18370 (lazy session creation):

1. _finalize_session() uses stale session_key after compression (#20001)
2. session_key not synced after auto-compression in run_conversation (#20001)
3. pending_title ValueError leaves title wedged forever (#19029)
4. Gateway silently swallows null responses when agent did work (#18765)
5. One-time cleanup for accumulated ghost compression continuations (#20001)

Changes:
- tui_gateway/server.py: _finalize_session() now uses agent.session_id
  (falls back to session_key when agent is None). Refactor
  _sync_session_key_after_compress() with clear_pending_title and
  restart_slash_worker policy flags. Call it post-run_conversation()
  to sync session_key after auto-compression. Add ValueError handler
  to pending_title flush.
- gateway/run.py: Extract _normalize_empty_agent_response() helper that
  consolidates failed/partial/null response handling. Surfaces user-facing
  error when agent did work (api_calls > 0) but returned no text.
- hermes_state.py: Add finalize_orphaned_compression_sessions() — marks
  ghost continuation sessions as ended (non-destructive, preserves data).
- cli.py: One-time startup migration for orphaned compression sessions.

Test changes:
- tests/test_tui_gateway_server.py: Update pending_title ValueError test
  for post-#18370 architecture (title applied post-message, not at create).
- tests/test_lazy_session_regressions.py: 14 new regression tests covering
  all fixed paths.

* docs(web_tools): correct web_extract summarizer timeout comment

The comment at tools/web_tools.py:700-702 stated the runtime default for
auxiliary.web_extract.timeout is 360s. The actual runtime default is 30s
(_DEFAULT_AUX_TIMEOUT in agent/auxiliary_client.py:3140), used by
_get_task_timeout when no auxiliary.web_extract.timeout key is present in
config.yaml.

The 360s figure is the config template default written by
hermes_cli/config.py:697 into freshly-generated config.yaml files. It only
takes effect when that key exists in the user's config — not as a fallback.
Users on configs that predate commit 20b4060d (Apr 5, 2026), or who removed
the key, fall through to the 30s _DEFAULT_AUX_TIMEOUT runtime default.

The comment was introduced in 20b4060d alongside the template-default bump
from 30 to 360. The runtime default in auxiliary_client.py was not changed
in that commit and has remained 30s since 839d9d74 (Mar 28, 2026).

* docs(config): fix fallback provider config paths

* docs(prompt): clarify supported customization surfaces

* chore: AUTHOR_MAP entry for Beandon13

* docs: remove dead reference links in flash-attention skill

* docs: remove dead papers.md link from saelens references

* docs: fix broken nix-setup anchor for container-aware CLI

* fix(telegram): keep DM topic typing scoped

* refactor(telegram): make typing thread-id resolver symmetric with send

Mirror _message_thread_id_for_typing() with _message_thread_id_for_send():
both now map the General forum topic (thread id "1") to None upfront.

That removes the need for the retry-without-thread fallback in send_typing()
entirely — if _message_thread_id_for_typing() returns a non-None value, it's
a real user-created topic and falling back to the root chat is never correct.
If Telegram rejects the typing action (e.g. topic deleted mid-session), we
swallow it at debug level instead of bleeding the indicator into All Messages.

Updates the General-topic typing regression test to assert the new single-call
contract.

* docs(tts): document per-provider max_text_length caps

PR #13743 replaced the global MAX_TEXT_LENGTH=4000 with a per-provider
table and a user-override 'max_text_length:' key, but the user-guide
TTS page documented no length behaviour at all. Users hitting truncation
had no way to discover the new caps or the override.

Add an 'Input length limits' subsection after the existing Configuration
YAML block: provider default caps (Edge 5000 / OpenAI 4096 / xAI 15000 /
MiniMax 10000 / Mistral 4000 / Gemini 5000 / ElevenLabs model-aware /
NeuTTS,KittenTTS 2000), ElevenLabs model_id -> cap table (5k-40k), an
override example, and the validation rules (non-positive / non-integer /
boolean values fall through to the provider default).

* docs(skill/hermes-agent): sync slash commands + add durable-systems section

Mirrors the AGENTS.md #20226 additions (Toolsets / Delegation / Curator /
Cron / Kanban) into the user-facing hermes-agent skill, and closes the
drift in the in-session slash command list.

User report (wxrrior in Discord): the skill did not mention /goal, so a
brand-new session answering "/hermes-agent do you have any info on /goal"
confidently said it did not exist. Cross-check against the CommandDef
registry found 16 commands missing from the static list: /goal, /agents,
/busy, /copy, /curator, /debug, /footer, /gquota, /indicator, /kanban,
/redraw, /reload, /reload-skills, /snapshot, /steer, /topic.

Changes:
- Slash Commands header now tells the reader to run /help or check the
  live docs reference as the source of truth, and names the registry
  of record (hermes_cli/commands.py) so future drift gets flagged
  honestly instead of answered confidently wrong.
- Added all 16 missing commands, slotted into existing subsections
  (/goal and /steer in Session; /busy + /indicator + /footer in
  Configuration; /curator + /kanban + /reload-skills + /reload in
  Tools & Skills; /topic in Gateway; /copy in Utility; /gquota +
  /debug in Info).
- Toolsets table updated to the authoritative 30-key list from
  toolsets.py (added kanban, yuanbao, spotify, safe, debugging, video,
  feishu_doc, feishu_drive, discord, discord_admin, clarify; previously
  stopped at 20 keys).
- New "Durable & Background Systems" section before Troubleshooting
  covers Delegation, Cron, Curator, Kanban - each with a short rundown
  of CLI verbs, key invariants, and a pointer to the user-facing docs.
  Mirrors AGENTS.md #20226 but in the skill's user-facing register.
- Bumped version 2.0.0 -> 2.1.0.

* docs(cli): add --deliver-only flag to hermes webhook subscribe

PR #12473 (merged 2026-04-19) added a new --deliver-only flag to
`hermes webhook subscribe` for zero-LLM direct delivery, but
website/docs/reference/cli-commands.md options table did not
reference it. Add the row so CLI users can discover the flag from
the reference page instead of having to read the source.

* perf(ui-tui): narrow overlay subscriptions to focused selectors

Subscribe overlay components to computed theme/session selectors instead of the full UI store so unrelated UI state updates trigger fewer overlay renders.

* docs(cli): add skills reset subcommand to CLI reference

PR #11468 added `hermes skills reset` but cli-commands.md was not
updated. Adds the subcommand to the table and usage examples.

Closes #11543

* feat(kanban): generic diagnostics engine for task distress signals (#20332)

* feat(kanban): generic diagnostics engine for task distress signals

Replaces the hallucination-specific ``warnings`` / ``RecoverySection``
surface (shipped in PR #20232) with a reusable diagnostic-rule engine
that covers five distress kinds in v1 and can be extended without
touching UI code. The "something's wrong with this task" signal is
no longer limited to phantom card ids.

Closes the follow-up from #20232 discussion.

New module
----------
``hermes_cli/kanban_diagnostics.py`` — stateless, no-side-effect rule
engine. Each rule is a pure function of
``(task, events, runs, now, config) -> list[Diagnostic]``. Registry
is a simple list; adding a new distress kind is one function + one
import, no UI or API changes required.

v1 rule set
-----------
* ``hallucinated_cards`` (error) — folds the existing
  ``completion_blocked_hallucination`` event into the new surface.
* ``prose_phantom_refs`` (warning) — folds
  ``suspected_hallucinated_references``.
* ``repeated_spawn_failures`` (error → critical at 2x threshold) —
  fires when ``tasks.spawn_failures >= 3``; suggests
  ``hermes -p <profile> doctor`` / ``auth``.
* ``repeated_crashes`` (error → critical) — fires after N consecutive
  ``crashed`` run outcomes with no successful completion between;
  suggests ``hermes kanban log <id>``.
* ``stuck_in_blocked`` (warning) — fires after 24h in ``blocked``
  state with no comments / unblock attempts; suggests commenting.

Every diagnostic carries structured ``actions`` (reclaim, reassign,
unblock, cli_hint, comment, open_docs) that render consistently in
both CLI and dashboard. Suggested actions are highlighted; generic
recovery actions (reclaim / reassign) are available on every kind as
fallbacks.

Diagnostics auto-clear when the underlying failure resolves — a
clean ``completed``/``edited`` event drops hallucination diagnostics,
a successful run drops crash diagnostics, a comment drops
stuck-blocked diagnostics. Audit events persist; the badge goes away.

API
---
``plugin_api.py``:
* ``/board`` now attaches ``diagnostics`` (full list) and
  ``warnings`` (compact summary with ``highest_severity``) per task.
* ``/tasks/{id}`` attaches diagnostics so the drawer's Diagnostics
  section auto-opens on flagged tasks.
* NEW ``/diagnostics`` endpoint — fleet-wide listing, filterable by
  severity, sorted critical-first.

CLI
---
* NEW ``hermes kanban diagnostics [--severity X] [--task id]
  [--json]`` — fleet view or single-task view, matches dashboard rule
  output so CLI users see the same picture.
* ``hermes kanban show <id>`` now renders a Diagnostics section near
  the top with severity markers + suggested actions.

Dashboard
---------
* Card badge is severity-coloured (⚠ amber warning, !! orange error,
  !!! red critical) using ``warnings.highest_severity``.
* Attention strip above the toolbar counts EVERY task with active
  diagnostics (not just hallucinations), severity-coloured, lists
  affected tasks with Open buttons when expanded.
* Drawer's old ``RecoverySection`` replaced with generic
  ``DiagnosticsSection`` rendering a card per active diagnostic:
  title + detail + structured data (task-id chips when payload keys
  look like id lists) + action buttons. Reassign profile picker is
  inline per-diagnostic. Clipboard fallback uses ``.catch()`` for
  environments where writeText rejects.
* Three-rung severity palette; amber for warning, orange for error,
  red for critical. Uses CSS variables so theming is straightforward.

Tests
-----
* NEW ``tests/hermes_cli/test_kanban_diagnostics.py`` — 14 unit tests
  covering each rule's positive/negative/threshold paths, severity
  sorting, broken-rule isolation, and sqlite3.Row integration.
* Dashboard plugin tests extended: ``/diagnostics`` endpoint (empty,
  populated, severity-filtered), ``/board`` exposes both diagnostic
  list and compact summary with ``highest_severity``.
* Existing hallucination-specific test (``test_board_surfaces_
  warnings_field_for_hallucinated_completions``) updated to reflect
  the new contract: warning summary keys by diagnostic kind
  (``hallucinated_cards``) not event kind.

379 kanban-suite tests pass (+16 net from this PR).

Live verification
-----------------
Seeded all 5 diagnostic kinds + one clean + one plain-running task
(7 total) into an isolated HERMES_HOME, spun up the dashboard, and
verified:

* Attention strip: shows ``!! 5 tasks need attention`` in the
  error-severity orange; Show expands to a list of 5 rows ordered
  critical > error > warning.
* Card badges: error tasks render ``!!`` orange, warning tasks
  render ``⚠`` amber, clean and plain-running tasks render no badge.
* Each of the 5 rules opens a correctly-coloured, correctly-styled
  diagnostic card in the drawer with its specific suggested action.
* Live reassign from a diagnostic card flipped
  ``broken-ml-worker → alice`` and the drawer refreshed with the
  new assignee + the same diagnostic still firing (correct:
  spawn_failures counter hasn't reset yet).
* CLI ``hermes kanban diagnostics`` prints all 5 in severity order;
  ``--severity error`` narrows to 3; ``kanban show <id>`` includes
  the Diagnostics block at the top with suggested action hint.

Migration note
--------------
The old ``warnings`` shape (``{count, kinds, latest_at}``) is
preserved on the API but ``kinds`` now keys by diagnostic kind
(``hallucinated_cards``) instead of event kind
(``completion_blocked_hallucination``). ``highest_severity`` is a
new required field. The dashboard was the only consumer and has
been updated in the same commit; external API consumers of the
``warnings`` field will need to update their kind-match logic.

* feat(kanban/diagnostics): lead titles with the actual error text

The generic 'Worker crashed N runs in a row' / 'Worker failed to spawn
N times' titles buried the actual cause in the data section. Operators
had to open logs or expand the diagnostic to see WHY the worker is
stuck — rate-limit vs insufficient quota vs bad auth vs context
overflow vs network blip all looked identical at a glance.

New titles:

  Agent crashed 3x: openai: 429 Too Many Requests - rate limit reached
  Agent crashed 3x: anthropic: 402 insufficient_quota - credit balance
  Agent crashed 3x: provider auth error: 401 Unauthorized
  Agent spawn failed 4x: insufficient_quota: You exceeded your current

Detail keeps the full error snippet (capped at 500 chars + ellipsis
for tracebacks). Title takes the first line capped at 160 chars.
Fallback title if no error recorded stays honest ('no error recorded').

Tests: 4 new cases covering 429/billing/spawn/truncation. 383 total
pass (+4).

Live-verified on dashboard with 6 seeded scenarios
(rate-limit, billing, auth, context, network, spawn-billing) —
each card title leads with the actionable error text.

* docs(agent): remove stale BuiltinMemoryProvider references from memory module docstrings

The BuiltinMemoryProvider class was removed from the codebase but its
name lingered in the module-level docstrings of memory_manager.py and
memory_provider.py, creating false expectations:

- memory_manager.py docstring showed example code doing
  add_provider(BuiltinMemoryProvider(...)) which ImportError at runtime
- memory_provider.py docstring listed BuiltinMemoryProvider as
  'always present, not removable' — misleading for new contributors

The regression test (test_memory_user_id.py) already passes without
any reference to BuiltinMemoryProvider; it uses RecordingProvider
instances directly. The stale references were docs-only drift.

Update both docstrings to reflect the actual current architecture:
MemoryManager accepts external plugin providers only (one at a time).

Closes #14402

* docs(plugins): document ctx.dispatch_tool() in plugin capabilities table

* docs(guide): add Dispatch tools from slash commands section

* docs(cron): add context_from chaining section

Resolved merge against current main (new No-agent mode section added in parallel).

Co-authored-by: Tony Simons <tony@tonysimons.dev>

* chore: AUTHOR_MAP entry for asimons81

* feat: provider modules — ProviderProfile ABC, 33 providers, fetch_models, transport single-path

Introduces providers/ package — single source of truth for every
inference provider. Adding a simple api-key provider now requires one
providers/<name>.py file with zero edits anywhere else.

What this PR ships:
- providers/ package (ProviderProfile ABC + 33 profiles across 4 api_modes)
- ProviderProfile declarative fields: name, api_mode, aliases, display_name,
  env_vars, base_url, models_url, auth_type, fallback_models, hostname,
  default_headers, fixed_temperature, default_max_tokens, default_aux_model
- 4 overridable hooks: prepare_messages, build_extra_body,
  build_api_kwargs_extras, fetch_models
- chat_completions.build_kwargs: profile path via _build_kwargs_from_profile,
  legacy flag path retained for lmstudio/tencent-tokenhub (which have
  session-aware reasoning probing that doesn't map cleanly to hooks yet)
- run_agent.py: profile path for all registered providers; legacy path
  variable scoping fixed (all flags defined before branching)
- Auto-wires: auth.PROVIDER_REGISTRY, models.CANONICAL_PROVIDERS,
  doctor health checks, config.OPTIONAL_ENV_VARS, model_metadata._URL_TO_PROVIDER
- GeminiProfile: thinking_config translation (native + openai-compat nested)
- New tests/providers/ (79 tests covering profile declarations, transport
  parity, hook overrides, e2e kwargs assembly)

Deltas vs original PR (salvaged onto current main):
- Added profiles: alibaba-coding-plan, azure-foundry, minimax-oauth
  (were added to main since original PR)
- Skipped profiles: lmstudio, tencent-tokenhub stay on legacy path (their
  reasoning_effort probing has no clean hook equivalent yet)
- Removed lmstudio alias from custom profile (it's a separate provider now)
- Skipped openrouter/custom from PROVIDER_REGISTRY auto-extension
  (resolve_provider special-cases them; adding breaks runtime resolution)
- runtime_provider: profile.api_mode only as fallback when URL detection
  finds nothing (was breaking minimax /v1 override)
- Preserved main's legacy-path improvements: deepseek reasoning_content
  preserve, gemini Gemma skip, OpenRouter response caching, Anthropic 1M
  beta recovery, etc.
- Kept agent/copilot_acp_client.py in place (rejected PR's relocation —
  main has 7 fixes landed since; relocation would revert them)
- _API_KEY_PROVIDER_AUX_MODELS alias kept for backward compat with existing
  test imports

Co-authored-by: kshitijk4poor <82637225+kshitijk4poor@users.noreply.github.com>
Closes #14418

* feat(providers): make all 33 providers pluggable under plugins/model-providers/

Every provider profile is now a self-contained plugin under
plugins/model-providers/<name>/, mirroring the plugins/platforms/
pattern established for IRC and Teams. The ProviderProfile ABC
stays in providers/; the per-provider profile data moves out.

- plugins/model-providers/<name>/__init__.py calls register_provider()
- plugins/model-providers/<name>/plugin.yaml declares kind: model-provider
- providers/__init__.py._discover_providers() lazily scans bundled plugins
  then $HERMES_HOME/plugins/model-providers/<name>/ (user override path)
- User plugins with the same name override bundled ones (last-writer-wins
  in register_provider)
- Legacy providers/<name>.py layout still supported for back-compat with
  out-of-tree editable installs
- Hermes PluginManager: new kind=model-provider; skipped like memory
  plugins (providers/ discovery owns them); standalone plugins with
  register_provider+ProviderProfile in their __init__.py auto-coerce to
  this kind (same heuristic as memory providers)
- skip_names extended to include 'model-providers' so the general
  PluginManager doesn't double-scan the category
- 4 new tests in tests/providers/test_plugin_discovery.py covering
  bundled discovery, user override, and general-loader isolation
- Docs updated: website/docs/developer-guide/adding-providers.md,
  provider-runtime.md, providers/README.md, plugins/model-providers/README.md

No API break: auth.py / config.py / doctor.py / models.py / runtime_provider.py /
model_metadata.py / auxiliary_client.py / chat_completions.py / run_agent.py
all still consume providers via get_provider_profile() / list_providers() —
they just now see plugin-discovered entries instead of pkgutil-iterated ones.

Third parties can now drop a single directory into
~/.hermes/plugins/model-providers/<name>/ to add or override an inference
provider without touching the repo.

* docs(cli): expand hermes import reference — add description, warning, and examples

* docs(bedrock): fix IAM permissions, add quickstart entry, add fallback provider, fix deployment section

* docs: fix Camofox Docker setup instructions

* docs(providers): Together/Groq/Perplexity cookbook via custom_providers

Three worked recipes for OpenAI-compatible cloud providers, plus the
Copilot HTTP 401 auto-recovery info block and the GMI Cloud row in the
compatible providers table. All three additions were on the original
docs/custom-providers-cookbook branch but its merge base predated 1186
main commits, making the rebase impractical (84k+ line conflict).

Replays just the providers.md additions onto current main.

* fix(tui): close slash parity gaps with CLI (#20339)

* fix(tui): close slash parity gaps with CLI

Route unsupported /skills subcommands through slash.exec, support /new <name>
titles, and handle /redraw natively so TUI behavior matches classic CLI. Also
filter gateway-only commands out of the TUI catalog while keeping /status
discoverable.

* fix(tui): run remaining CLI parity paths natively

Forward chat launch flags into the TUI runtime and handle live-session status
and skill reloads in the gateway process so TUI state no longer depends on the
slash worker's stale CLI instance.

* fix(tui): block stale snapshot restores

Prevent snapshot restore from running through the isolated slash worker because
it mutates disk state without refreshing the live TUI agent.

* chore: uptick

* fix(tui): guard async session title updates

Handle failures from the fire-and-forget session.title RPC so title-setting errors do not surface as unhandled promise rejections while preserving session-scoped messaging.

* docs(gemini): add Google Gemini guide

* chore: AUTHOR_MAP entry for jethac

* docs: align terminal-backend count and naming across docs and code

README:24 claimed "Six terminal backends" while tools/environments/ exposes
seven top-level backend choices through TERMINAL_ENV: local, docker, ssh,
singularity, modal, daytona, vercel_sandbox. Modal additionally has direct
and Nous-managed modes selected via terminal.modal_mode (the
ManagedModalEnvironment class is a Modal sub-mode, not a separate top-level
backend).

The same drift appeared in five other doc and code-comment sites with
inconsistent counts (six, seven, or implicit) and varying lists. Updated
all sites to a consistent seven-backend list in canonical order. The
configuration guide also clarifies how Modal's two modes are selected so
operators do not search for a non-existent backend: managed_modal value.

CONTRIBUTING.md:160 lists six backend filenames in a code tree but does
not carry the "Six terminal" prose; left out of scope per cohesion sweep
guidance to bundle only identical wording.

Files updated:
- README.md (line 24, marketing copy)
- website/docs/index.md (line 49, landing page)
- website/docs/user-guide/configuration.md (line 86, config guide)
- tools/environments/__init__.py (lines 3-6, package docstring)
- tools/file_operations.py (line 6, module docstring)
- environments/README.md (line 43, RL training docs — TERMINAL_ENV list)

* chore: AUTHOR_MAP entry for deep-name

* docs: refresh stale platform/LOC/test counts; clarify gateway vs plugin platforms

AGENTS.md is the AI-assistant entry doc, so its counts get used as ground
truth. Several values had drifted, and the same drift had spread to a few
user-facing surfaces. Fixing all of them in one commit so the count claims
agree and clearly distinguish gateway-core from plugin-shipped platforms.

AGENTS.md:
- run_agent.py "~12k LOC" → "~14k LOC as of 2026-05-03" (actual 14,097)
- cli.py     "~11k LOC" → "~12k LOC as of 2026-05-03" (actual 12,043)
- tools/environments/ list now lists all 7 user-selectable terminal backends
  in canonical order, matching tools/terminal_tool.py:2214-2215
- gateway/platforms/ list adds yuanbao and wecom_callback; the 19 names
  match the user-facing list at website/docs/integrations/index.md
- plugins/ tree now mentions plugins/platforms/ (irc, teams)
- tests/ snapshot "~15k tests across ~700 files as of Apr 2026" →
  "~19k tests across ~890 files as of 2026-05-03"

User-facing count claims:
- hermes_cli/tips.py:195 — "19 platforms" → "21 messaging platforms" with
  IRC and Microsoft Teams added to the named list
- website/docs/index.md:49 — "6 terminal backends" → "7 terminal backends:
  ..., Vercel Sandbox" (also corrected by PR #19044; same edit content)
- website/docs/index.md:50 — "15+ platforms from one gateway" → "21+ messaging
  platforms (19 in the gateway, plus IRC and Microsoft Teams via plugins)"
- website/docs/integrations/index.md:83-85 — "15+ messaging platforms" → "19+",
  added yuanbao to the linked list. The surrounding text scopes it to "configured
  through the same gateway subsystem", so plugin platforms (IRC, Teams) are
  intentionally not in this list
- website/scripts/generate-llms-txt.py:205 — "15+ platforms" → "21+ messaging
  platforms — 19 native to the gateway plus IRC and Microsoft Teams via plugins"

LOC and date stamps follow the existing AGENTS.md "as of <date>" convention
(line 56 already used this pattern). Source of truth for the gateway count is
gateway/config.py:130-148 (PlatformID enum); plugin platforms live in
plugins/platforms/.

Out of scope:
- RELEASE_v0.9.0.md historical "16 platforms" claim (immutable history)
- userStories.json verbatim user quotes
- Programmatic count generation from gateway/config.py + plugin manifests
  is a worthwhile build-system change but separate from these content fixes

* docs(skills): explain restoring bundled skills

* docs(docker): add section on connecting to local inference servers (vLLM, Ollama)

Adds a comprehensive guide for connecting Dockerized Hermes to local
inference servers like vLLM and Ollama, covering:
- Docker Compose networking (recommended)
- Standalone Docker run with host.docker.internal / --network host
- Connectivity verification steps
- Ollama-specific example

Closes #12308

* docs(docker): document API_SERVER_* env vars for exposing the OpenAI-compatible endpoint

Salvage of #11758. The PR's original diff was stale (the Docker Compose section on main has been heavily refactored — dashboard is now an embedded side-process, not a separate service), so the useful bit (API server env var requirements) is applied as a note on the basic `docker run` example.

Co-authored-by: xiangyong <xiangyong@zspace.cn>

* chore: AUTHOR_MAP entry for CES4751

* docs(discord): fix Server Members Intent + SSRC-mapping drift; add /voice join slash Choice

Salvage of #11350. Kept:
- Code: add an explicit /voice join Choice in the slash UI (runner accepts both 'join' and 'channel' but only 'channel' was in autocomplete).
- Docs: Server Members Intent is conditional (only needed if DISCORD_ALLOWED_USERS contains usernames); SSRC → user_id mapping uses the voice websocket SPEAKING opcode, not the Members intent.

Dropped from the original PR:
- HERMES_DISCORD_VOICE_PACKET_DUMP — this env var doesn't exist on main (it was in a different PR that isn't merged).
- DISCORD_PROXY docs — already documented on current main.
- DISCORD_ALLOW_MENTION_* docs — already on main.
- "barge-in mode" rewrite — current main actually does pause the listener during TTS (VoiceReceiver.pause() at discord.py:192); there is no barge_in_guard/barge_in_rms on main.

Co-authored-by: Michel Belleau <michel.belleau@malaiwah.com>

* docs(skills): modernize Obsidian file workflows

* chore: AUTHOR_MAP entry for counterposition

* docs(kanban): document handoff evidence metadata

* chore: AUTHOR_MAP entry for Fearvox

* docs: clarify Telegram group chat troubleshooting

* docs(codex): clarify OAuth auth prerequisite

* docs(voice): add Doubao speech integration examples (TTS + STT)

* chore: AUTHOR_MAP entry for Hypnus-Yuan

* docs(faq): use messaging extra for gateway deps

* chore: AUTHOR_MAP entry for xsfX20

* fix(kanban): unify failure counter across spawn/timeout/crash outcomes (#20410)

The dispatcher's circuit breaker only protected against spawn-side
failures (profile missing, workspace mount error, exec failure).
Workers that successfully spawned but then timed out or crashed
re-queued to ``ready`` with no counter increment, so the next tick
re-spawned them — loops forever until someone noticed. Reported
externally on Twitter (Forbidden Seeds) and confirmed by walking the
kernel: ``enforce_max_runtime`` flipped the task back to ready, emitted
a ``timed_out`` event, and never touched ``spawn_failures``; same for
``detect_crashed_workers``.

Fix: unify the counter across all non-success outcomes.

Schema
------
* ``tasks.spawn_failures`` → ``tasks.consecutive_failures``
* ``tasks.last_spawn_error`` → ``tasks.last_failure_error``
* Migration renames the columns in-place on existing DBs (``ALTER
  TABLE RENAME COLUMN`` — SQLite >= 3.25) so historical counter
  values are preserved. Row mappers fall through to the legacy names
  if both column renames and a migration somehow got out of sync.

Counter lifecycle
-----------------
New helper ``_record_task_failure(conn, task_id, error, *, outcome,
release_claim, end_run, event_payload_extra)`` is the single point
every non-success outcome funnels through:

* ``spawn_failed``  → ``_record_spawn_failure`` (kept as alias)
  calls it with ``release_claim=True, end_run=True`` — transitions
  running→ready, clears claim, closes run.
* ``timed_out`` → ``enforce_max_runtime`` already does the status
  transition + run close + event emission, then calls
  ``_record_task_failure`` with ``release_claim=False, end_run=False``
  just to bump the counter (and trip the breaker if needed).
* ``crashed`` → ``detect_crashed_workers`` same pattern, but the
  counter increment runs after the main write_txn closes (SQLite
  doesn't nest write transactions).

If the counter hits the breaker threshold (``DEFAULT_FAILURE_LIMIT=5``,
same as before), the task transitions to ``blocked`` with a ``gave_up``
event on top of whatever outcome-specific event was already emitted.

Reset semantics changed: the counter now clears only on successful
``complete_task`` (and operator ``reclaim_task`` — an explicit "I've
looked at this, try again with a fresh budget"). Previously
``_clear_spawn_failures`` ran on every successful spawn, which would
have wiped the counter before a timeout could accumulate past threshold
— exactly the loop this fix prevents.

Diagnostics
-----------
* ``_rule_repeated_spawn_failures`` → ``_rule_repeated_failures``. Now
  fires regardless of which outcome is at fault. Classifies the most
  recent failure (spawn_failed / timed_out / crashed) from the run
  history so the title ("Agent timeout x3", "Agent crash x4", "Agent
  spawn x5") and suggested action (``doctor`` for spawn, ``log`` for
  timeout/crash) stay outcome-specific without N duplicate rules.
* ``_rule_repeated_crashes`` kept as a narrower early-warning at
  threshold 2 (vs 3 for the unified rule), but now suppresses itself
  when the unified rule would also fire — avoids double-flagging.
* Diagnostic ``data`` payload now carries
  ``{consecutive_failures, most_recent_outcome, last_error}`` instead
  of spawn-specific keys.

CLI
---
* ``Task.consecutive_failures`` / ``Task.last_failure_error`` are the
  public fields now. Existing callers that referenced the old names
  get migrated (tests updated in this commit).
* Backward-compat: ``DEFAULT_SPAWN_FAILURE_LIMIT``,
  ``_clear_spawn_failures``, ``_record_spawn_failure`` stay as aliases.

Tests
-----
* 6 new kernel tests: timeout increments counter, 3 consecutive
  timeouts trip the breaker (was the reported gap), crash increments
  counter, reclaim clears counter, completion clears counter, spawn
  success does NOT clear counter.
* Diagnostic tests: updated ``repeated_spawn_failures`` cases to use
  the new kind name and add a timeout-loop test.
* Dashboard API test: spawn_failures column update → consecutive_failures.

389/389 kanban-suite tests pass.

Live verification
-----------------
Seeded 4 tasks in an isolated HERMES_HOME: 3 timeouts, 4 crashes,
2-spawn-failed + 2-timed-out, and a task that had prior failures but
completed successfully. Board correctly shows "!! 3 tasks need
attention" (the successful one has no badge because the counter
reset). Drawer for the timeout-loop task renders "Agent timeout x3"
with most_recent_outcome=timed_out and the "Check logs" suggested
action (not the spawn-flavoured "Verify profile"). The successful
task has zero diagnostics.

Closes the Forbidden-Seeds-reported gap.

* docs(guides): add guide for running Hermes locally with Ollama

Step-by-step guide covering Ollama installation, model selection,
Hermes configuration, speed optimization, and optional gateway bot
setup — all running on local hardware with zero API cost.

Includes hardware requirements, model comparison table with tool-call
support status, context window tuning, GPU offloading tips, fallback
provider setup, troubleshooting, and cost comparison.

* chore: AUTHOR_MAP entry for binhnt92

* docs: add Open WebUI bootstrap script

* chore: AUTHOR_MAP entry for acesjohnny

* docs(browser): document WSL-to-Windows Chrome MCP bridge

* chore: AUTHOR_MAP entry for liu-collab

* docs(i18n): add zh-Hans Tool Gateway, image gen, and Windows WSL guide

Made-with: Cursor

* docs: add Chinese (zh-CN) README translation

Closes #12954

- Add README.zh-CN.md with complete Simplified Chinese translation
- Add language switcher badge in README.md linking to Chinese version
- Add language switcher badge in README.zh-CN.md linking to English version

* chore: AUTHOR_MAP entry for zhanggttry

* docs: update VS Code setup instructions for ACP Client integration

* chore: AUTHOR_MAP entry for formulahendry

* test(kanban): cover metadata handoff round-trip

* feat(gateway): respect kanban.max_spawn config to limit concurrent tasks

The dispatch_once function already accepts a max_spawn parameter but the
gateway was calling it without passing any value, effectively ignoring
the configuration. This change reads kanban.max_spawn from config.yaml
and passes it through, allowing users to limit concurrent kanban tasks.

This prevents resource exhaustion scenarios where kanban dispatcher
spawns too many parallel workers on constrained hardware.

* guard kanban worker lifecycle by run id

* chore(release): AUTHOR_MAP entries for momowind and misery-hl

* feat(hindsight): probe API for update_mode='append' support, dedupe across processes

Mirrors the pattern already shipping in hindsight-integrations/openclaw:
probe `<api_url>/version` once per process, gate on Hindsight ≥ 0.5.0.
When supported, retains use a stable session-scoped `document_id`
(`session_id`) plus `update_mode='append'` so cross-process retains for
the same session merge into one document instead of producing
N-different-process-stamped duplicates. When unsupported (or probe
fails), fall back to the existing per-process unique
`f"{session_id}-{start_ts}"` document_id with no `update_mode` — the
resume-overwrite fix (#6654) keeps working unchanged on legacy servers.

Closes the dedup half of #20115. The proposed `document_id_strategy`
config knob isn't needed: auto-detection via the same /version probe
the OpenClaw plugin already uses gives the same outcome with no extra
config burden, and the choice is purely a function of what the server
can do.

Plumbing
--------
- Module-level helpers (`_meets_minimum_version`, `_fetch_hindsight_api_version`,
  `_check_api_supports_update_mode_append`) cache the result per api_url
  so every provider in the process gets one /version round-trip.
- One-time WARN logged when the API is older than 0.5.0, telling the
  user to upgrade for cross-session deduplication.
- New instance helper `_resolve_retain_target(fallback_doc_id)` returns
  `(document_id, update_mode)` based on cached capability. Wired into
  `sync_turn` and the `on_session_switch` flush path.
- For local_embedded mode, the probe URL is taken from the running
  client (`client.url`) so we hit the actual daemon port rather than
  the configured default.
- `update_mode` is set on the per-item dict; `aretain_batch` already
  threads `item['update_mode']` into the API call.

Tests
-----
- `TestUpdateModeAppendCapability` (5 cases): legacy fallback, modern
  stable+append, per-url cache, one-time warn, flush-on-switch resolves
  against the OLD session.
- Existing `_make_hindsight_provider` factory in the manager-side test
  file extended to seed `_mode`/`_api_url`/`_api_key`/`_client` and stub
  `_resolve_retain_target` so the bypass-init pattern keeps working.

E2E verified against installed `~/.hermes/hermes-agent`:
- Legacy probe (unreachable host) → `legacy-session-<ts>` doc_id,
  no `update_mode`.
- Modern probe (live local_embedded 0.5.6 daemon) → stable
  `modern-session` doc_id + `update_mode='append'`.
- `test_hermes_embedded_smoke.py` passes (90s).

* fix(api_server): SSE token batching + error handling for Open WebUI performance

Reduces SSE event rate ~500/turn → ~20/turn via 50ms text-delta batching in
_dispatch(), which eliminates markdown re-render storms on Open WebUI. Also:

- Trim tool_call.arguments in the response.completed event to 100KB
  (prevents silent hangs on 848KB+ single-line SSE events).
- Catch-all exception handlers in _write_sse_responses() + _write_sse_chat_completion()
  emit a proper error chunk instead of TransferEncodingError from incomplete
  chunked encoding when the agent crashes mid-stream.
- MAX_REQUEST_BYTES 1MB → 10MB; pass client_max_size to aiohttp Application to
  avoid silent 400s on truncated request bodies for long conversations.

Salvage of #17552 (api_server portion only). The contrib/openwebui-filter/
payload from that PR — Open WebUI Filter Function + benchmark writeup — is
a client-side user-installable add-on and doesn't need to live in the repo;
dropped here. Closes #17537.

Co-authored-by: bogerman1 <93757150+bogerman1@users.noreply.github.com>

* chore: AUTHOR_MAP entry for bogerman1

* feat(i18n): add French (fr) locale support

- Add fr.yaml with French translations for approval prompts and gateway messages
- Register 'fr' in SUPPORTED_LANGUAGES
- Add French aliases: french, français, fr-fr, fr-be, fr-ca, fr-ch
- Update locale sync comment in en.yaml

* feat(i18n): add Ukrainian locale

* chore: AUTHOR_MAP entry for olisikh

* arcee temperature + compression

* test(arcee): cover Trinity Large Thinking temperature + compression overrides

Salvage follow-up for PR #20344:
- AUTHOR_MAP entry for rob-maron (required by CI)
- 17 parametrized tests covering _is_arcee_trinity_thinking,
  _fixed_temperature_for_model Trinity override, and
  _compression_threshold_for_model, including sibling-model negatives
  (trinity-large-preview, trinity-mini) and the OpenRouter slug form.

* fix(doctor): report Kanban worker tools as runtime-gated

* fix(kanban): accept created_cards linked as child of completing task

Widens _verify_created_cards to also accept ids that are children of the
completing task in task_links. Previously we only accepted cards where
created_by matched the completing task's assignee, which was too strict
for legitimate orchestrator flows: a specifier creates a card (so
created_by=specifier, not worker), then a worker picks it up and passes
parents=[current_task] to kanban_create. The explicit link proves the
relationship and should be trusted.

Salvaged from #20022 @LeonSGP43 (full PR superseded by #20232 +
this patch; the linked-children relaxation was the portable
improvement).

* fix(kanban): measure max runtime from current run

* test(kanban): backdate task_runs.started_at alongside tasks.started_at

After #19473 landed (enforce_max_runtime reads from task_runs.started_at
rather than tasks.started_at), a regression test added earlier still
only backdated the tasks column. Backdate both so the test is robust
regardless of which column the enforcer reads from.

* fix(kanban): prevent child task dispatch when parent is not done

Add parent dependency guard to _set_status_direct so dragging
a task to the ready column is rejected (409) when its parents
are not all done. Previously the guard only existed in
recompute_ready, allowing direct status writes via the
dashboard API to bypass the dependency engine.

Root cause: after reclaiming stale workers, both T3 and T4
were set to ready via dashboard status writes in quick
succession, causing the writer to be spawned while the analyst
was blocked — upstream work wasn't done yet.

* feat(kanban): surface task_runs.summary on dashboard cards + ``kanban show``

The kanban-worker skill (built into the gateway dispatcher's spawn
prompt) instructs every worker to hand off via
``kanban_complete(summary=..., metadata=...)``. That writes the summary
onto the closing ``task_runs`` row, NOT onto ``tasks.result`` — the
latter is left NULL unless the caller passes ``result=`` explicitly.

Result: a glance at the dashboard or ``hermes kanban show <id>`` shows
a blank "Result:" section even when the worker did real work, which
on 2026-05-05 caused a Mac false-alarm ("Hermes did nothing") on a
task that had a 10-line completion summary on its run.

This patch surfaces the latest non-null run summary as
``latest_summary`` so the worker's actual handoff lands in front of
operators.

* New helpers ``kanban_db.latest_summary(conn, task_id)`` and
  ``kanban_db.latest_summaries(conn, task_ids)``. The batch variant
  uses a single window-function SELECT so the dashboard board endpoint
  doesn't pay an N+1 cost on multi-hundred-task boards.
* CLI ``hermes kanban show <id>`` prints a "Latest summary:" block
  when ``tasks.result`` is empty but a run has produced a summary
  (the existing "Result:" section still wins when populated, so the
  back-compat path for hand-edited results is untouched). JSON output
  gains a top-level ``latest_summary`` field.
* Dashboard ``/board`` and ``/tasks/{id}`` now include a
  ``latest_summary`` field on every task. Cards on /board carry a
  200-character preview (cheap to render, plenty for "what did this
  worker do?" at a glance); the drawer/detail endpoint returns the
  full summary.
* Five new tests cover: empty-runs case, post-complete surface,
  newest-of-multiple selection, empty-string skip, batch with
  missing tasks + empty input.

Smoke-tested locally against the live profile DB on the three
acceptance-criterion targets (t_f08fef91 cron-hygiene-audit,
t_007b7f1c EMA-analysis, t_05746fa4 self-assessment) — all three now
return their populated summaries via both ``latest_summary`` and
``latest_summaries``.

Test plan: 255/255 kanban tests pass + 91/91 dashboard plugin tests
pass. No regression on tasks where ``tasks.result`` is explicitly
populated (the existing "Result:" branch is preserved).

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

* fix(kanban): wire dependency selects

* chore(release): AUTHOR_MAP entries for suncokret12 and mioimotoai-lgtm

* feat(i18n): add Turkish (tr) locale

- Add locales/tr.yaml with Turkish translations for all approval.* and gateway.* keys
- Register 'tr' in SUPPORTED_LANGUAGES
- Add Turkish aliases: turkish, türkçe, tr-tr

* fix: add Turkish locale references in config, tests, and docs

- hermes_cli/config.py: add tr to supported languages comment
- locales/en.yaml: add tr to locale file list comment
- tests/agent/test_i18n.py: add Turkish alias tests + explicit lang test
- website/docs/user-guide/configuration.md: add tr to supported values

* docs: document custom model aliases for /model command (#20475)

User-defined model aliases (config.yaml model_aliases: and
model.aliases.*) have worked since early versions but were entirely
undocumented. Add a dedicated 'Custom model aliases' section to
slash-commands.md covering both YAML config formats and the
'hermes config set' shell form, mirror a shorter version into the
configuring-models 'Alternative methods' section, and cross-link from
the two /model table rows.

Flagged by @weehowe on Twitter — he wasn't aware the feature existed.

* feat(models): add deepseek/deepseek-v4-pro to OpenRouter + Nous Portal curated lists (#20495)

Endpoint re-tested over 6 conversational turns (9 API calls, 3 tool calls)
and an 8-request burst — no rate limits, no errors, ~2-3s latency. The
historical rate-limit issues that caused its removal are gone.

- hermes_cli/models.py: add to OPENROUTER_MODELS and _PROVIDER_MODELS['nous']
- website/static/api/model-catalog.json: regenerated via build_model_catalog.py

* feat(models): add x-ai/grok-4.3 to OpenRouter + Nous Portal curated lists (#20497)

Endpoint validated over 6 conversational turns with tool calls (9 API
calls, 3 tool calls, 0 failures) and an 8-request burst (8/8 ok,
0 rate limits). Latency ~5-10s/call — slower than grok-4.20 but
expected for a reasoning model.

- hermes_cli/models.py: add to OPENROUTER_MODELS and _PROVIDER_MODELS['nous']
- website/static/api/model-catalog.json: regenerated

* fix: salvage batch — compaction guidance, memory authority, cache eviction after compression

- Fix /compact → /compress in context-overflow tips (closes #20020)
- Evict cached agent after session hygiene and /compress so system
  prompt refreshes with current SOUL.md, memory, and skills
- Restore memory authority across compaction: change 'informational
  background data' to 'authoritative reference data' in memory block
  and SUMMARY_PREFIX, with backward-compatible regex

Based on:
- PR #20027 by @LeonSGP43
- PR #18767 by @MacroAnarchy
- PR #17380 by @vominh1919

PR #17121 boundary marker fix already merged to main (2eef395e1).
PR #9262 user-message anchoring already on main via _ensure_last_user_message_in_tail().

* feat(browser): add Lightpanda engine support with automatic Chrome fallback

Add Lightpanda as an optional browser engine for local mode.
Lightpanda is a headless browser built from scratch in Zig -- faster
navigation than Chrome with significantly less memory.

One config line to enable:
  browser:
    engine: lightpanda

New functions in browser_tool.py:
- _get_browser_engine() -- config/env reader with validation + caching
- _should_inject_engine() -- only inject in local non-cloud mode
- _needs_lightpanda_fallback() -- detect empty/failed LP results
- _chrome_fallback_screenshot() -- temporary Chrome session for screenshots
- Engine injection in _run_browser_command (--engine flag)
- browser_vision pre-routes screenshots to Chrome when engine=lightpanda

Config:
- browser.engine in DEFAULT_CONFIG (auto/lightpanda/chrome)
- AGENT_BROWSER_ENGINE in OPTIONAL_ENV_VARS
- /browser status shows engine info in local mode

Rebased from PR #7144 onto current main. All existing code preserved --
pure additions only (+520/-2).

25 new tests + 81 total browser tests pass (0 failures).

* fix(browser): surface Lightpanda Chrome fallback warnings

* feat(tui): collapsible sections in startup banner (skills, system prompt, MCP)

The TUI SessionPanel banner now uses collapsible \u25b8/\u25be toggle
sections matching the existing Chevron convention used for runtime
agent details. Skills, system prompt, and MCP server lists are
collapsed by default; tools remain expanded as the most actionable
info.

- tui_gateway/server.py: _session_info() now passes agent._cached_system_prompt
  through to the TUI frontend
- ui-tui/src/types.ts: added system_prompt?: string to SessionInfo
- ui-tui/src/components/branding.tsx: rewrote SessionPanel with
  CollapseToggle helper + per-section useState toggles

Default states: tools=open, skills=collapsed, system=collapsed,
mcp=collapsed. Clicking any \u25b8/\u25be header toggles that section.

* fix(tui): collapse long system messages in transcript with expand toggle

System messages over 400 chars (system prompt, AGENTS.md, etc.) now
render as a collapsed \u25b8/\u25be toggle line in the transcript, matching
the Chevron convention used for runtime details. The summary shows
the first line + char count; clicking expands to full content.

* fix(browser): tighten Lightpanda fallback edge cases

* fix(gateway): preserve model picker current context

* fix(update): drop pip --quiet so slow installs don't look hung (#20679)

On Termux/Android aarch64 (and other platforms without prebuilt wheels
for some optional extras), 'pip install -e .[all]' compiles C/Rust
extensions from source. This can run for several minutes with zero
network activity and — with --quiet — zero stdout. Users report
'hermes update hangs at Updating Python dependencies', Ctrl+C it, then
re-run and see 'up to date' (because git pull already succeeded and the
pip step was still working when they interrupted).

Pip's default output is proportional to actual work (one line per
Collecting / Building wheel for X / Installing), so removing --quiet
costs nothing on fast hardware and prevents the false-hang interrupt
loop on slow hardware.

Reported via Discord on Termux/Android. Supersedes #20466 which
misdiagnosed the hang as PYTHONPATH shadowing (install.sh doesn't run
during 'hermes update', and terminal() doesn't inherit PYTHONPATH).

* fix(cli): guard logger.debug in signal handler (#13710 regression) (#20673)

CPython's logging module is not reentrant-safe.  `Logger.isEnabledFor`
caches level results in `Logger._cache`; under shutdown races the cache
can be cleared (`Logger._clear_cache`, triggered by logging config changes
from another thread) or mid-mutation when a signal fires, raising
`KeyError: <level_int>` (e.g. `KeyError: 10` for DEBUG) inside the signal
handler.

When that happens, the KeyError escapes before the `raise KeyboardInterrupt()`
on the next line can fire, which bypasses prompt_toolkit's normal interrupt
unwind and surfaces as the EIO cascade originally reported in #13710.

Issue #13710 shipped two defenses (asyncio exception handler + outer
`except (KeyError, OSError)` with EIO suppression) that cover the EIO
unwind path.  This patch closes the remaining escape hatch: the
`logger.debug` call at the top of `_signal_handler` itself.  Wrap it in a
bare `try/except Exception: pass` so logging can never raise through a
signal handler.

Observed in the wild: debug report on 0.12.0 (commit 8163d371) shows the
exact stack — KeyError: 10 at logging/__init__.py:1742 inside the
signal handler's `logger.debug`, followed by the EIO cascade from
prompt_toolkit's emergency flush.

Tests: adds `TestSignalHandlerLoggingRace` to
`tests/hermes_cli/test_suppress_eio_on_interrupt.py` with 6 new cases:
- normal path still raises KeyboardInterrupt
- KeyError(10) from logger.debug does not escape
- any Exception from logger.debug is swallowed
- agent.interrupt still fires when logger.debug raises
- agent.interrupt raising also does not escape
- BaseException (SystemExit) is NOT swallowed — guard uses `except Exception`
  deliberately so real shutdown signals still propagate

Closes #13710 regression.

* fix: harden install.sh against inherited Python env leakage

* chore: AUTHOR_MAP entry for adybag14-cyber

* fix(ui): reduce status-line jitter while scrolling

* fix(tui): stabilize FaceTicker elapsed width to prevent composer drift

* fix(tui): restore gap before duration when verb segment is hidden

The verb-padding change dropped the leading space in durationSegment on
the assumption that the verb's trailing pad always supplies the gap. But
the unicode spinner style sets showVerb=false, making verbSegment an
empty string — in that mode the output would become `{frame}· {duration}`
with no separator. Add the space back; harmless when the verb segment
is shown (its trailing pad still provides the gap).

* chore(release): map liuguangyong@hellobike -> liuguangyong93

* fix(kanban): reset code element background inside board

The Nous DS globals.css applies a global rule:
  code { background: var(--midground); color: var(--background); }

This paints an opaque cream/yellow fill on every <code> element,
which hides text in the kanban drawer's event-payload, run-meta,
and worker-log panes (all rendered as <code>).

Fix: scope a reset inside .hermes-kanban so <code> elements inherit
their parent's color and stay transparent.

* fix(cli): recover classic CLI output after resize

* feat(skills): add shop-app personal shopping assistant (optional) (#20702)

Port Shop.app's upstream SKILL.md (https://shop.app/SKILL.md) into
optional-skills/productivity/shop-app/ with Hermes-native adaptations:

- Proper Hermes frontmatter (name, description<=60 chars, version,
  author, license, prerequisites, metadata.hermes tags + related_skills
  + homepage + upstream)
- Swap Shop.app's bespoke 'message()' tool references for Hermes
  conventions: gateway adapters handle platform formatting, so the
  skill just writes markdown (no Telegram/WhatsApp/iMessage sections
  referencing a tool Hermes doesn't ship)
- Name Hermes tools where relevant: curl via 'terminal', HTML policy
  pages via 'web_extract', try-on via 'image_generate'
- Reframe session state as 'hold in your reasoning context for this
  conversation only' and forbid writing tokens to .env / disk — matches
  Hermes ephemeral-memory discipline
- Drop NO_REPLY convention (Shop-app-runtime specific)
- Trigger-first description so the skill loader picks it up when the
  user wants to search products, track orders, returns, or reorder

* feat(checkpoints): v2 single-store rewrite with real pruning + disk guardrails (#20709)

Replaces the per-directory shadow-repo design with a single shared shadow
git store at ~/.hermes/checkpoints/store/. Object DB is now deduplicated
across every working directory the agent has ever touched; a dozen
worktrees of the same project cost near-zero in additional disk.

Why
---
Pre-v2 design had three compounding problems that let ~/.hermes/checkpoints/
grow to multi-GB on active machines:

1. Each working directory got its own full shadow git repo — no object
   dedup across projects or across worktrees of the same project.
2. _prune() was a documented no-op: max_snapshots only limited the
   /rollback listing. Loose objects accumulated forever.
3. Defaults: enabled=True, auto_prune=False — users paid the disk cost
   without ever asking for /rollback.

Field report on a single workstation: 847 MB across 47 shadow repos,
mostly redundant clones of the hermes-agent source tree.

Changes
-------
- tools/checkpoint_manager.py: full rewrite. Single bare store, per-project
  refs (refs/hermes/<hash>), per-project indexes (store/indexes/<hash>),
  per-project metadata (store/projects/<hash>.json with workdir +
  created_at + last_touch). On first v2 init, any pre-v2 per-directory
  shadow repos are auto-migrated into legacy-<timestamp>/ so the new
  store starts clean. _prune() now actually rewrites the per-project ref
  to the last max_snapshots commits and runs git gc --prune=now. New
  _enforce_size_cap() drops oldest commits round-robin across projects
  when the store exceeds max_total_size_mb. _drop_oversize_from_index()
  filters any single file larger than max_file_size_mb out of the snapshot.
- hermes_cli/checkpoints.py: new 'hermes checkpoints' CLI
  (status / list / prune / clear / clear-legacy) for managing the store
  outside a session.
- hermes_cli/config.py: flipped defaults — enabled=False, max_snapshots=20,
  auto_prune=True. Added max_total_size_mb=500, max_file_size_mb=10.
  Tightened DEFAULT_EXCLUDES (added target/, *.so/*.dylib/*.dll,
  *.mp4/*.mov, *.zip/*.tar.gz, .worktrees/, .mypy_cache/, etc.).
- run_agent.py / cli.py / gateway/run.py: thread the new kwargs through
  AIAgent and the startup auto_prune hooks.
- Tests rewritten to match v2 storage while keeping backwards-compat
  coverage for the pre-v2 prune path (per-directory shadow repos under
  base/ are still swept correctly for anyone mid-migration).
- Docs updated: user-guide/checkpoints-and-rollback.md explains the
  shared store, new defaults, migration, and the new CLI;
  reference/cli-commands.md documents 'hermes checkpoints'.

E2E validated
-------------
- Legacy migration: pre-v2 shadow repos auto-archived into legacy-<ts>/.
- Object dedup: two projects with an identical shared.py blob resolve to
  7 total objects in the store (v1 would have stored the blob twice).
- max_snapshots=3 actually enforced: after 6 commits, list shows 3.
- Orphan prune: deleting a project's workdir + 'hermes checkpoints prune
  --retention-days 0' removes its ref, index, and metadata; GC reclaims
  the objects.
- max_file_size_mb=1 excludes a 2 MB weights.bin while keeping the
  tracked source code files.
- hermes checkpoints {status,prune,clear,clear-legacy} all work from the
  CLI without an agent running.

Breaking / migration
--------------------
No in-place data migration — legacy per-directory shadow repos are moved
into legacy-<timestamp>/ on first run. Old /rollback history is still
accessible by inspecting the archive with git; run
'hermes checkpoints clear-legacy' to reclaim the space when ready. Users
relying on /rollback must now set checkpoints.enabled=true (or pass
--checkpoints) explicitly.

* fix(cli): catch OSError in _resolve_attachment_path to prevent ENAMETOOLONG dropping long slash commands

When the user pastes a long slash command like \`/goal <long prose>\` into
\`hermes chat\`, the input flows into \`_detect_file_drop()\`, whose
\`starts_like_path\` prefilter accepts anything starting with \`/\` and
forwards it to \`_resolve_attachment_path()\`. That helper calls
\`Path.exists()\` which invokes \`os.stat()\`, which raises
\`OSError(errno=ENAMETOOLONG)\` — 63 on macOS, 36 on Linux — when the
candidate exceeds NAME_MAX (typically 255 bytes).

The OSError propagates up to the broad \`except Exception\` in
\`process_loop\` (cli.py:11798), gets logged at WARNING level, and the
user's input is silently dropped. From the user's POV the chat prompt
hangs — the only signal is in agent.log:

  WARNING cli: process_loop unhandled error (msg may be lost):
    [Errno 63] File name too long: "/goal Drive the space board..."

This affects any slash command with prose-length arguments — \`/goal\`
in particular but also \`/skill\`, \`/cron\`, custom user commands.

Fix: wrap the \`exists()\`/\`is_file()\` calls in try/except OSError so
structurally-invalid path candidates cleanly return None. The slash-
command dispatch path downstream (cli.py:11718) then handles the
input correctly.

Tests: two new regression cases in test_cli_file_drop.py cover the
original \`/goal\` reproducer and a synthetic long path. All 35 file-
drop tests pass.

Reproducer (without the fix):
  python -c "from cli import _detect_file_drop;
             _detect_file_drop('/goal ' + 'a'*300)"
  → OSError: [Errno 63] File name too long

* chore(release): map cleo@edaphic.xyz → curiouscleo

Follow-up to the salvaged fix for /goal ENAMETOOLONG drop — adds
AUTHOR_MAP entry so the release script resolves the commit author to
the correct GitHub user.

* docs(wsl2): expand Windows (WSL2) guide — filesystem, networking, services, pitfalls (#20748)

Replaces the 22-line stub with a ~320-line guide covering the parts of the
Windows/WSL2 split that specifically affect Hermes users:

- Why WSL2 (and not native Windows)
- Install: distro choice, WSL1→2, systemd via /etc/wsl.conf
- Filesystem boundary: /mnt/c vs \\wsl$, perf/perms/watchers/case,
  wslpath/wslview, CRLF + git core.autocrlf, clone-where guidance
- Networking in both directions:
  - WSL → Windows services: links to the canonical WSL2 Networking section
    in integrations/providers.md (mirrored mode, NAT + host IP, bind addr,
    firewall) instead of duplicating
  - Windows/LAN → Hermes in WSL: mirrored vs NAT, netsh portproxy one-liner,
    firewall rule, webhook tunneling pointer
- Long-running services: systemd gateway + Task Scheduler wsl.exe --exec
  'sleep infinity' to keep the VM alive at login
- GPU passthrough: NVIDIA works, AMD/Intel out of matrix
- Common pitfalls: connection refused, /mnt/c slowness, CRLF ^M,
  UNC warnings, post-sleep clock drift, mirrored-mode DNS with VPN,
  PATH, Defender scanning, VHDX disk reclaim

All internal links use site-absolute /docs/... form (matches the rest of
user-guide/); all seven link targets verified to exist.

* docs: pluggable surfaces coverage — model-provider guide, full plugin map, opt-in fix (#20749)

* docs(providers): add model-provider-plugin authoring guide + fix stale refs

New docs:
- website/docs/developer-guide/model-provider-plugin.md — full authoring
  guide (directory layout, minimal example, ProviderProfile fields,
  overridable hooks, user overrides, api_mode selection, auth types,
  testing, pip distribution)
- Wired into website/sidebars.ts under 'Extending'
- Cross-references added in:
  - guides/build-a-hermes-plugin.md (tip block)
  - developer-guide/adding-providers.md
  - developer-guide/provider-runtime.md

User guide:
- user-guide/features/plugins.md: Plugin types table grows from 3 to 4
  with 'Model providers' row

Stale comment cleanup (providers/*.py → plugins/model-providers/<name>/):
- hermes_cli/main.py:_is_profile_api_key_provider docstring
- hermes_cli/doctor.py:_build_apikey_providers_list docstring
- hermes_cli/auth.py: PROVIDER_REGISTRY + alias auto-extension comments
- hermes_cli/models.py: CANONICAL_PROVIDERS auto-extension comment

AGENTS.md:
- Project-structure tree: added plugins/model-providers/ row
- New section: 'Model-provider plugins' explaining discovery, override
  semantics, PluginManager integration, kind auto-coerce heuristic

Verified: docusaurus build succeeds, new page renders, all 3 cross-links
resolve. 347/347 targeted tests pass (tests/providers/,
tests/hermes_cli/test_plugins.py, tests/hermes_cli/test_runtime_provider_resolution.py,
tests/run_agent/test_provider_parity.py).

* docs(plugins): add 'pluggable interfaces at a glance' maps to plugins.md + build-a-hermes-plugin

Devs landing on either the user-guide plugin page or the build-a-plugin
guide now get an upfront table of every distinct pluggable surface with
a link to the right authoring doc. Previously they'd have to read the
full general-plugin guide to discover that model providers / platforms
/ memory / context engines are separate systems.

user-guide/features/plugins.md:
- New 'Pluggable interfaces — where to go for each' section below the
  existing…

PR NousResearch#13743 replaced the global MAX_TEXT_LENGTH=4000 with a per-provider
table and a user-override 'max_text_length:' key, but the user-guide
TTS page documented no length behaviour at all. Users hitting truncation
had no way to discover the new caps or the override.

Add an 'Input length limits' subsection after the existing Configuration
YAML block: provider default caps (Edge 5000 / OpenAI 4096 / xAI 15000 /
MiniMax 10000 / Mistral 4000 / Gemini 5000 / ElevenLabs model-aware /
NeuTTS,KittenTTS 2000), ElevenLabs model_id -> cap table (5k-40k), an
override example, and the validation rules (non-positive / non-integer /
boolean values fall through to the provider default).

PR NousResearch#13743 replaced the global MAX_TEXT_LENGTH=4000 with a per-provider
table and a user-override 'max_text_length:' key, but the user-guide
TTS page documented no length behaviour at all. Users hitting truncation
had no way to discover the new caps or the override.

Add an 'Input length limits' subsection after the existing Configuration
YAML block: provider default caps (Edge 5000 / OpenAI 4096 / xAI 15000 /
MiniMax 10000 / Mistral 4000 / Gemini 5000 / ElevenLabs model-aware /
NeuTTS,KittenTTS 2000), ElevenLabs model_id -> cap table (5k-40k), an
override example, and the validation rules (non-positive / non-integer /
boolean values fall through to the provider default).

PR NousResearch#13743 replaced the global MAX_TEXT_LENGTH=4000 with a per-provider
table and a user-override 'max_text_length:' key, but the user-guide
TTS page documented no length behaviour at all. Users hitting truncation
had no way to discover the new caps or the override.

Add an 'Input length limits' subsection after the existing Configuration
YAML block: provider default caps (Edge 5000 / OpenAI 4096 / xAI 15000 /
MiniMax 10000 / Mistral 4000 / Gemini 5000 / ElevenLabs model-aware /
NeuTTS,KittenTTS 2000), ElevenLabs model_id -> cap table (5k-40k), an
override example, and the validation rules (non-positive / non-integer /
boolean values fall through to the provider default).

PR NousResearch#13743 replaced the global MAX_TEXT_LENGTH=4000 with a per-provider
table and a user-override 'max_text_length:' key, but the user-guide
TTS page documented no length behaviour at all. Users hitting truncation
had no way to discover the new caps or the override.

Add an 'Input length limits' subsection after the existing Configuration
YAML block: provider default caps (Edge 5000 / OpenAI 4096 / xAI 15000 /
MiniMax 10000 / Mistral 4000 / Gemini 5000 / ElevenLabs model-aware /
NeuTTS,KittenTTS 2000), ElevenLabs model_id -> cap table (5k-40k), an
override example, and the validation rules (non-positive / non-integer /
boolean values fall through to the provider default).

NousResearch#13743)

A single global MAX_TEXT_LENGTH = 4000 truncated every TTS provider at
4000 chars, causing long inputs to be silently chopped even though the
underlying APIs allow much more:

  - OpenAI:     4096
  - xAI:        15000
  - MiniMax:    10000
  - ElevenLabs: 5000 / 10000 / 30000 / 40000 (model-aware)
  - Gemini:     ~5000
  - Edge:       ~5000

The schema description also told the model 'Keep under 4000 characters',
which encouraged the agent to self-chunk long briefs into multiple TTS
calls (producing 3 separate audio files instead of one).

New behavior:
  - PROVIDER_MAX_TEXT_LENGTH table + ELEVENLABS_MODEL_MAX_TEXT_LENGTH
    encode the documented per-provider limits.
  - _resolve_max_text_length(provider, cfg) resolves:
      1. tts.<provider>.max_text_length user override
      2. ElevenLabs model_id lookup
      3. provider default
      4. 4000 fallback
  - text_to_speech_tool() and stream_tts_to_speaker() both call the
    resolver; old MAX_TEXT_LENGTH alias kept for back-compat.
  - Schema description no longer hardcodes 4000.

Tests: 27 new unit + E2E tests; all 53 existing TTS tests and 253
voice-command/voice-cli tests still pass.

PR NousResearch#13743 replaced the global MAX_TEXT_LENGTH=4000 with a per-provider
table and a user-override 'max_text_length:' key, but the user-guide
TTS page documented no length behaviour at all. Users hitting truncation
had no way to discover the new caps or the override.

Add an 'Input length limits' subsection after the existing Configuration
YAML block: provider default caps (Edge 5000 / OpenAI 4096 / xAI 15000 /
MiniMax 10000 / Mistral 4000 / Gemini 5000 / ElevenLabs model-aware /
NeuTTS,KittenTTS 2000), ElevenLabs model_id -> cap table (5k-40k), an
override example, and the validation rules (non-positive / non-integer /
boolean values fall through to the provider default).

NousResearch#13743)

A single global MAX_TEXT_LENGTH = 4000 truncated every TTS provider at
4000 chars, causing long inputs to be silently chopped even though the
underlying APIs allow much more:

  - OpenAI:     4096
  - xAI:        15000
  - MiniMax:    10000
  - ElevenLabs: 5000 / 10000 / 30000 / 40000 (model-aware)
  - Gemini:     ~5000
  - Edge:       ~5000

The schema description also told the model 'Keep under 4000 characters',
which encouraged the agent to self-chunk long briefs into multiple TTS
calls (producing 3 separate audio files instead of one).

New behavior:
  - PROVIDER_MAX_TEXT_LENGTH table + ELEVENLABS_MODEL_MAX_TEXT_LENGTH
    encode the documented per-provider limits.
  - _resolve_max_text_length(provider, cfg) resolves:
      1. tts.<provider>.max_text_length user override
      2. ElevenLabs model_id lookup
      3. provider default
      4. 4000 fallback
  - text_to_speech_tool() and stream_tts_to_speaker() both call the
    resolver; old MAX_TEXT_LENGTH alias kept for back-compat.
  - Schema description no longer hardcodes 4000.

Tests: 27 new unit + E2E tests; all 53 existing TTS tests and 253
voice-command/voice-cli tests still pass.

PR NousResearch#13743 replaced the global MAX_TEXT_LENGTH=4000 with a per-provider
table and a user-override 'max_text_length:' key, but the user-guide
TTS page documented no length behaviour at all. Users hitting truncation
had no way to discover the new caps or the override.

Add an 'Input length limits' subsection after the existing Configuration
YAML block: provider default caps (Edge 5000 / OpenAI 4096 / xAI 15000 /
MiniMax 10000 / Mistral 4000 / Gemini 5000 / ElevenLabs model-aware /
NeuTTS,KittenTTS 2000), ElevenLabs model_id -> cap table (5k-40k), an
override example, and the validation rules (non-positive / non-integer /
boolean values fall through to the provider default).

NousResearch#13743)

A single global MAX_TEXT_LENGTH = 4000 truncated every TTS provider at
4000 chars, causing long inputs to be silently chopped even though the
underlying APIs allow much more:

  - OpenAI:     4096
  - xAI:        15000
  - MiniMax:    10000
  - ElevenLabs: 5000 / 10000 / 30000 / 40000 (model-aware)
  - Gemini:     ~5000
  - Edge:       ~5000

The schema description also told the model 'Keep under 4000 characters',
which encouraged the agent to self-chunk long briefs into multiple TTS
calls (producing 3 separate audio files instead of one).

New behavior:
  - PROVIDER_MAX_TEXT_LENGTH table + ELEVENLABS_MODEL_MAX_TEXT_LENGTH
    encode the documented per-provider limits.
  - _resolve_max_text_length(provider, cfg) resolves:
      1. tts.<provider>.max_text_length user override
      2. ElevenLabs model_id lookup
      3. provider default
      4. 4000 fallback
  - text_to_speech_tool() and stream_tts_to_speaker() both call the
    resolver; old MAX_TEXT_LENGTH alias kept for back-compat.
  - Schema description no longer hardcodes 4000.

Tests: 27 new unit + E2E tests; all 53 existing TTS tests and 253
voice-command/voice-cli tests still pass.

PR NousResearch#13743 replaced the global MAX_TEXT_LENGTH=4000 with a per-provider
table and a user-override 'max_text_length:' key, but the user-guide
TTS page documented no length behaviour at all. Users hitting truncation
had no way to discover the new caps or the override.

Add an 'Input length limits' subsection after the existing Configuration
YAML block: provider default caps (Edge 5000 / OpenAI 4096 / xAI 15000 /
MiniMax 10000 / Mistral 4000 / Gemini 5000 / ElevenLabs model-aware /
NeuTTS,KittenTTS 2000), ElevenLabs model_id -> cap table (5k-40k), an
override example, and the validation rules (non-positive / non-integer /
boolean values fall through to the provider default).