… never sent (#2336 Area 7) (#2471)

Follow-up to #2336 (umbrella) / #2463 (Areas 1+2+4). Completes Area 7 for
the Web UI surface only. Native apps (Android / iOS / macOS) deferred —
state-drivers there require release coordination; classification report
in PR description will feed that work.

## Classification

Web UI `thinkingLevel` was dead display plumbing: read from `chat.history`
into `state.chatThinkingLevel`, passed through `ChatProps.thinkingLevel`
to `renderChat`, but the view body never read the prop. No UI code
mutates or sends `thinkingLevel` — `patchSession` in `controllers/sessions.ts`
only handles `label` / `verboseLevel`.

## Changes

- `ui/src/ui/types.ts` — drop phantom `thinkingLevel?` from `GatewaySessionRow`
  and `SessionsPatchResult.entry`.
- `ui/src/ui/controllers/chat.ts` — drop `chatThinkingLevel` from `ChatState`,
  narrow `chat.history` response type, drop the state assignment.
- `ui/src/ui/app-view-state.ts`, `ui/src/ui/app.ts`, `ui/src/ui/app-render.ts` —
  drop `chatThinkingLevel` state plumbing.
- `ui/src/ui/views/chat.ts` — drop unused `thinkingLevel` prop from
  `ChatProps`. `renderChat` reads `activeSession?.reasoningLevel` for
  its single reasoning branch — not `thinkingLevel`.
- Tests: drop fixture/assertion lines that were exercising the removed
  display plumbing.

## Backward compat

Backend still sends `thinkingLevel` in `chat.history` response (until
#2464 removes `SessionEntry.thinkingLevel`). UI ignoring unknown response
fields is safe — correct ordering: reader removed first (this PR),
backend schema removal next (#2464).

## What was NOT touched — and why

- `ui/src/ui/format.test.ts` — tests `stripThinkingTags` (XML `<think>`
  tag stripping from CLI assistant output). Unrelated to `thinkingLevel`
  state field. Keep.
- `apps/shared/RemoteClawKit/Tests/OpenClawKitTests/AssistantTextParserTests.swift` —
  same pattern: `<think>` XML parsing. Keep.
- All native-app files (Android / iOS / macOS) — state-drivers that send
  `thinkingLevel` via `sessions.patch` and `chat.send(thinking:)`.
  Deferred per issue instruction; full classification in PR description.

## Verification

- `grep thinkingLevel|chatThinkingLevel|thinkLevel` in `ui/` → 0 matches.
- `pnpm check` → exit 0 (format + tsgo + lint + lint:tmp:no-random-messaging
  + lint:no-remoteclaw-ai).
- `pnpm test src/ui/controllers/chat.test.ts src/ui/views/chat.test.ts` →
  35/35 passed. Other UI test failures on this branch match `origin/main`
  baseline (unrelated to `thinkingLevel`).

Closes #2467. Unblocks #2464.

Refs: #2336, #2463, #2464

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

Part of the umbrella thinking-level / model-selection sweep (#2336).
Sibling PR #2463 landed Areas 1+2+4 (runtime `/think` parsing, helpers,
and passthrough plumbing); this PR covers Area 6 (docs).

RemoteClaw does not control thinking levels — CLI runtimes (Claude,
Gemini, Codex, OpenCode) own that decision. User-facing documentation
advertising `/think` as a directive was inherited from upstream
OpenClaw and is now stale.

## Removals

- `docs/tools/slash-commands.md`:
  - L17: drop `/think` from the directives list.
  - L114: delete the `/think <off|minimal|low|medium|high|xhigh>`
    command entry.
  - L133: drop `/think: high` from the colon-separator example
    (`/send: on`, `/help:` still illustrate both syntactic variants).
- `docs/concepts/context.md:139`: drop `/think` from the directives
  list.
- `docs/web/tui.md:94`: delete the `/think <off|...>` TUI entry.
- `docs/channels/group-messages.md`:
  - L18: drop `or /think high` (kept `/verbose on` as primary
    example).
  - L70: drop `/think high` from the session-level directives list.

## Verification

- `git grep "/think\b" docs/` → 0 hits (matches parent #2336 AC).
- `cd docs && npm run build` → 296 pages built, no errors.
- `pnpm format:docs:check` → clean.
- `pnpm lint:docs` → no errors on edited files (pre-existing errors on
  unrelated files documented; identical on `origin/main`).

## Out of scope (tracked as informational for future #2336 waves)

- `/thinking off` advice at `docs/help/faq.md:2354` — alias of the
  removed directive.
- `/tools/thinking` URL references in `faq.md`, `messages.md`,
  `hubs.md` — pre-existing broken links; target page was gutted in an
  earlier wave.
- `docs/web/tui.md:40` status-line string `model + think/fast/...` —
  UI indicator shorthand, not a slash command.
- `src/agents/agent-helpers/errors.ts` still references `/think` in an
  error message — src cleanup belongs to Areas 1/2/4.

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

… (#2472)

Part of the umbrella thinking-level / model-selection sweep (#2336).
Sibling PR #2463 landed Areas 1+2+4 (runtime `/think` parsing, helpers,
and passthrough plumbing); this PR covers Area 6 (docs).

RemoteClaw does not control thinking levels — CLI runtimes (Claude,
Gemini, Codex, OpenCode) own that decision. User-facing documentation
advertising `/think` as a directive was inherited from upstream
OpenClaw and is now stale.

## Removals

- `docs/tools/slash-commands.md`:
  - L17: drop `/think` from the directives list.
  - L114: delete the `/think <off|minimal|low|medium|high|xhigh>`
    command entry.
  - L133: drop `/think: high` from the colon-separator example
    (`/send: on`, `/help:` still illustrate both syntactic variants).
- `docs/concepts/context.md:139`: drop `/think` from the directives
  list.
- `docs/web/tui.md:94`: delete the `/think <off|...>` TUI entry.
- `docs/channels/group-messages.md`:
  - L18: drop `or /think high` (kept `/verbose on` as primary
    example).
  - L70: drop `/think high` from the session-level directives list.

## Verification

- `git grep "/think\b" docs/` → 0 hits (matches parent #2336 AC).
- `cd docs && npm run build` → 296 pages built, no errors.
- `pnpm format:docs:check` → clean.
- `pnpm lint:docs` → no errors on edited files (pre-existing errors on
  unrelated files documented; identical on `origin/main`).

## Out of scope (tracked as informational for future #2336 waves)

- `/thinking off` advice at `docs/help/faq.md:2354` — alias of the
  removed directive.
- `/tools/thinking` URL references in `faq.md`, `messages.md`,
  `hubs.md` — pre-existing broken links; target page was gutted in an
  earlier wave.
- `docs/web/tui.md:40` status-line string `model + think/fast/...` —
  UI indicator shorthand, not a slash command.
- `src/agents/agent-helpers/errors.ts` still references `/think` in an
  error message — src cleanup belongs to Areas 1/2/4.

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

…ea 5 of #2336

Follow-up to PR #2463 (Areas 1+2+4 of #2336). Completes the directive-handling
portion of the sweep by removing model-state plumbing that was already empty/
trivial after #2335's `createModelSelectionState` removal.

Per the Middleware Boundary Principle: RemoteClaw does NOT control model
selection — CLI runtimes (Claude, Gemini, Codex, OpenCode) own those decisions.
`handleDirectiveOnly` and `applyInlineDirectivesFastLane` carried 10 fields
(`aliasIndex`, `allowedModelKeys`, `allowedModelCatalog`, `resetModelOverride`,
`defaultProvider`, `defaultModel`, `initialModelLabel`, `formatModelSwitchEvent`,
`provider`, `model`) that are now fully dead.

## Changes

- `src/auto-reply/reply/directive-handling.params.ts`: delete all 10 model-state
  fields from `HandleDirectiveOnlyCoreParams`; delete unused `surface?: string`
  from `HandleDirectiveOnlyParams`.
- `src/auto-reply/reply/directive-handling.impl.ts`: queue-validation `channel`
  semantic fix — switch from misused `params.provider` (AI provider, e.g.
  "anthropic") to `params.messageProviderKey ?? ""` (actual message channel,
  e.g. "telegram"). Aligns with `get-reply-run.ts:382` which already passes
  `sessionCtx.Provider` correctly. User-visible effect is limited to
  channel-specific queue config display when running `/queue` without args.
- `src/auto-reply/reply/directive-handling.fast-lane.ts`: drop model-state
  destructures; drop dead `sessionEntry.providerOverride`/`modelOverride`
  reads; simplify return type `{directiveAck?; provider; model}` →
  `{directiveAck?}`. *Note*: `providerOverride`/`modelOverride` ARE written
  by `src/gateway/sessions-patch.ts:213-214` and
  `src/agents/tools/session-status-tool.ts:254-255`, but their downstream
  consumption via this fast-lane path reaches only `isReasoningTagProvider`
  at `get-reply-run.ts:453` (`enforceFinalTag` for google/google-gemini-cli/
  minimax) — vestigial Pi-era plumbing per #2336's gut thesis.
- `src/auto-reply/reply/get-reply-directives-apply.ts`: delete inline empty-
  default `directiveModelState` compat shim (introduced in #2335); delete
  `aliasIndex`/`initialModelLabel`/`formatModelSwitchEvent` threading; inline
  single-call `createDirectiveHandlingBase` factory; drop `provider`/`model`
  from `ApplyDirectiveResult.continue` variant (pure pass-through of inputs
  post-gut).
- `src/auto-reply/reply/get-reply-directives.ts` (required cascade): delete
  dead `initialModelLabel`/`formatModelSwitchEvent` constructions and their
  two param passes (orphaned by the scope reduction in apply.ts); collapse
  `let provider`/`let model` to `const` (no longer reassigned).

## Deferred

- `resolveDefaultModel` stub in `directive-handling.ts` still returns
  `allowedModelKeys`/`allowedModelCatalog`/`resetModelOverride` — never read
  by callers. Separate cleanup wave.
- `aliasIndex` still threaded through `get-reply.ts` → `resolveReplyDirectives`
  (dead throughout). Separate cleanup wave.
- Adjacent Areas 3 (`SessionEntry.thinkingLevel`), 6 (docs), 7 (display audit),
  8 (test-layer hardening) of #2336 remain as separate follow-ups.

## Verification

- `pnpm check`: exit 0 (format + tsgo + oxlint + project-specific lints)
- `pnpm test`: 800 files / 7010 passed / 3 skipped — matches #2463 baseline
- Grep across the 4 target files: zero hits for all 10 gutted symbols

Refs: #2465, #2336, #2335, #2463

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

…ea 5 of #2336 (#2475)

Follow-up to PR #2463 (Areas 1+2+4 of #2336). Completes the directive-handling
portion of the sweep by removing model-state plumbing that was already empty/
trivial after #2335's `createModelSelectionState` removal.

Per the Middleware Boundary Principle: RemoteClaw does NOT control model
selection — CLI runtimes (Claude, Gemini, Codex, OpenCode) own those decisions.
`handleDirectiveOnly` and `applyInlineDirectivesFastLane` carried 10 fields
(`aliasIndex`, `allowedModelKeys`, `allowedModelCatalog`, `resetModelOverride`,
`defaultProvider`, `defaultModel`, `initialModelLabel`, `formatModelSwitchEvent`,
`provider`, `model`) that are now fully dead.

## Changes

- `src/auto-reply/reply/directive-handling.params.ts`: delete all 10 model-state
  fields from `HandleDirectiveOnlyCoreParams`; delete unused `surface?: string`
  from `HandleDirectiveOnlyParams`.
- `src/auto-reply/reply/directive-handling.impl.ts`: queue-validation `channel`
  semantic fix — switch from misused `params.provider` (AI provider, e.g.
  "anthropic") to `params.messageProviderKey ?? ""` (actual message channel,
  e.g. "telegram"). Aligns with `get-reply-run.ts:382` which already passes
  `sessionCtx.Provider` correctly. User-visible effect is limited to
  channel-specific queue config display when running `/queue` without args.
- `src/auto-reply/reply/directive-handling.fast-lane.ts`: drop model-state
  destructures; drop dead `sessionEntry.providerOverride`/`modelOverride`
  reads; simplify return type `{directiveAck?; provider; model}` →
  `{directiveAck?}`. *Note*: `providerOverride`/`modelOverride` ARE written
  by `src/gateway/sessions-patch.ts:213-214` and
  `src/agents/tools/session-status-tool.ts:254-255`, but their downstream
  consumption via this fast-lane path reaches only `isReasoningTagProvider`
  at `get-reply-run.ts:453` (`enforceFinalTag` for google/google-gemini-cli/
  minimax) — vestigial Pi-era plumbing per #2336's gut thesis.
- `src/auto-reply/reply/get-reply-directives-apply.ts`: delete inline empty-
  default `directiveModelState` compat shim (introduced in #2335); delete
  `aliasIndex`/`initialModelLabel`/`formatModelSwitchEvent` threading; inline
  single-call `createDirectiveHandlingBase` factory; drop `provider`/`model`
  from `ApplyDirectiveResult.continue` variant (pure pass-through of inputs
  post-gut).
- `src/auto-reply/reply/get-reply-directives.ts` (required cascade): delete
  dead `initialModelLabel`/`formatModelSwitchEvent` constructions and their
  two param passes (orphaned by the scope reduction in apply.ts); collapse
  `let provider`/`let model` to `const` (no longer reassigned).

## Deferred

- `resolveDefaultModel` stub in `directive-handling.ts` still returns
  `allowedModelKeys`/`allowedModelCatalog`/`resetModelOverride` — never read
  by callers. Separate cleanup wave.
- `aliasIndex` still threaded through `get-reply.ts` → `resolveReplyDirectives`
  (dead throughout). Separate cleanup wave.
- Adjacent Areas 3 (`SessionEntry.thinkingLevel`), 6 (docs), 7 (display audit),
  8 (test-layer hardening) of #2336 remain as separate follow-ups.

## Verification

- `pnpm check`: exit 0 (format + tsgo + oxlint + project-specific lints)
- `pnpm test`: 800 files / 7010 passed / 3 skipped — matches #2463 baseline
- Grep across the 4 target files: zero hits for all 10 gutted symbols

Refs: #2465, #2336, #2335, #2463

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

CLI runtimes own reasoning depth — the thinkingLevel input pipeline
has no runtime effect. Remove the remaining surface after #2463-#2471
eliminated most of it.

Removals:
- Delete orphan src/auto-reply/thinking.shared.ts (no importers)
- Prune dead helpers in thinking.ts: ThinkingCatalogEntry,
  isBinaryThinkingProvider, listThinkingLevels, shadow
  normalizeProviderId
- Drop subagents.thinking from agent-defaults types/schema and
  AgentEntry.subagents zod schema
- Remove --thinking flag from cron add/edit + tests; strip thinking
  from CronPayload, normalize, store-migration, and cron jobs
  mergePayload/buildPayloadFromPatch
- Remove thinking from gateway agent.send params, HookAgentPayload,
  HookAction, HookMappingResolved, server-http, server/hooks,
  server-node-events agent.request deep link
- Remove thinking from AgentCommandOpts + thinkingOnce,
  agent-via-gateway, subagent-spawn _thinkingOverrideRaw /
  thinkingOverride, sessions_spawn tool schema
- Remove thinking from protocol schemas (agent.ts, cron.ts,
  logs-chat.ts, sessions.ts) and ACP translator chat.send path
- Remove hooks.mappings[].thinking / hooks.gmail.thinking from zod
  schema, types, help, labels, quality test; regenerate
  schema.base.generated.ts
- Remove dead --thinking option + help-text example from agent CLI
  register; remove [--thinking] from /subagents spawn usage line;
  remove "thinking" from cron-tool agentTurn payload doc + JOB_KEYS
- Remove dead mocks (normalizeThinkLevel, supportsXHighThinking) and
  dead resolvedThinkLevel param in buildStatusReply
- Sweep docs: 9 files no longer teach --thinking / thinkingDefault

Added:
- LEGACY_CONFIG_RULES entries for agents.defaults.thinkingDefault and
  agents.defaults.subagents.thinking
- strip-thinking-level-fields LEGACY_CONFIG_MIGRATIONS entry that
  strips thinkingDefault, agents.defaults.subagents.thinking,
  agents.list[].subagents.thinking, hooks.mappings[].thinking, and
  hooks.gmail.thinking so existing configs still load cleanly
- 3 new tests in config-misc.test.ts covering the new migration

Kept (intentional):
- type === "thinking" / thinking_delta output-event rendering in
  middleware/runtimes, transcript truncation, ACP output blocks,
  TUI formatters, HTML export template
- Telegram/Discord "thinking" lifecycle status emoji (different
  concept from thinking level)
- reasoningLevel everywhere (different concept — channel-bound
  display behavior)

remoteclaw cron add --thinking high now errors with
"unknown option '--thinking'".

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

…) (#2486)

CLI runtimes own reasoning depth — the thinkingLevel input pipeline
has no runtime effect. Remove the remaining surface after #2463-#2471
eliminated most of it.

Removals:
- Delete orphan src/auto-reply/thinking.shared.ts (no importers)
- Prune dead helpers in thinking.ts: ThinkingCatalogEntry,
  isBinaryThinkingProvider, listThinkingLevels, shadow
  normalizeProviderId
- Drop subagents.thinking from agent-defaults types/schema and
  AgentEntry.subagents zod schema
- Remove --thinking flag from cron add/edit + tests; strip thinking
  from CronPayload, normalize, store-migration, and cron jobs
  mergePayload/buildPayloadFromPatch
- Remove thinking from gateway agent.send params, HookAgentPayload,
  HookAction, HookMappingResolved, server-http, server/hooks,
  server-node-events agent.request deep link
- Remove thinking from AgentCommandOpts + thinkingOnce,
  agent-via-gateway, subagent-spawn _thinkingOverrideRaw /
  thinkingOverride, sessions_spawn tool schema
- Remove thinking from protocol schemas (agent.ts, cron.ts,
  logs-chat.ts, sessions.ts) and ACP translator chat.send path
- Remove hooks.mappings[].thinking / hooks.gmail.thinking from zod
  schema, types, help, labels, quality test; regenerate
  schema.base.generated.ts
- Remove dead --thinking option + help-text example from agent CLI
  register; remove [--thinking] from /subagents spawn usage line;
  remove "thinking" from cron-tool agentTurn payload doc + JOB_KEYS
- Remove dead mocks (normalizeThinkLevel, supportsXHighThinking) and
  dead resolvedThinkLevel param in buildStatusReply
- Sweep docs: 9 files no longer teach --thinking / thinkingDefault

Added:
- LEGACY_CONFIG_RULES entries for agents.defaults.thinkingDefault and
  agents.defaults.subagents.thinking
- strip-thinking-level-fields LEGACY_CONFIG_MIGRATIONS entry that
  strips thinkingDefault, agents.defaults.subagents.thinking,
  agents.list[].subagents.thinking, hooks.mappings[].thinking, and
  hooks.gmail.thinking so existing configs still load cleanly
- 3 new tests in config-misc.test.ts covering the new migration

Kept (intentional):
- type === "thinking" / thinking_delta output-event rendering in
  middleware/runtimes, transcript truncation, ACP output blocks,
  TUI formatters, HTML export template
- Telegram/Discord "thinking" lifecycle status emoji (different
  concept from thinking level)
- reasoningLevel everywhere (different concept — channel-bound
  display behavior)

remoteclaw cron add --thinking high now errors with
"unknown option '--thinking'".

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