fix(agents): keep hook context in embedded prompts by Alix-007 · Pull Request #86875 · openclaw/openclaw

Alix-007 · 2026-05-26T11:43:47Z

Summary

keep before_prompt_build prependContext/appendContext in the embedded runner model prompt when transcriptPrompt is present
strip marked OpenClaw internal runtime-context blocks into hidden runtime context messages
fail closed on malformed/unterminated internal runtime-context blocks so hidden text cannot leak into the model prompt
update context-engine and prompt-splitter regression coverage for prompt-local hooks, multi-block extraction, inline marker mentions, and malformed blocks

Real behavior proof

Behavior or issue addressed: Embedded-runner before_prompt_build prependContext/appendContext must remain in the model prompt when a separate persisted transcriptPrompt exists, while valid and malformed internal runtime-context blocks stay out of the model prompt.
Real environment tested: Local OpenClaw checkout at PR head 1d17a3f5596c, Linux, Node v22.22.0, runtime module src/agents/pi-embedded-runner/run/runtime-context-prompt.ts executed through node --import tsx.
Exact steps or command run after this patch: Ran the real exported resolveRuntimeContextPromptParts() against four cases: valid hook context plus hidden block, unterminated hidden block, multiple hidden blocks, and repeated inline marker mentions.
Evidence after fix: Terminal output from the runtime command:

{
  "environment": "node v22.22.0 platform=linux",
  "head": "1d17a3f5596c",
  "validPrompt": "dynamic hook context\n\nvisible ask\n\ndynamic hook tail",
  "validRuntimeContext": "<<<BEGIN_OPENCLAW_INTERNAL_CONTEXT>>>\nsecret runtime context\n<<<END_OPENCLAW_INTERNAL_CONTEXT>>>",
  "validPromptLeaksSecret": false,
  "malformedPrompt": "visible ask",
  "malformedLeaksSecret": false,
  "multiRuntimeBlocks": 2,
  "inlinePromptHasMarkerMention": true,
  "inlinePromptLeaksSecret": false
}

Observed result after fix: Hook prefix/tail remained model-visible, valid hidden context moved to runtimeContext, unterminated hidden context was dropped fail-closed after the visible prompt, two hidden blocks were extracted, and inline marker mentions remained visible without leaking secret block content.
What was not tested: No browser/channel UI rendering was run manually; the integration regression covers raw transcript/UI text remaining separate from prompt-local hook context.

Verification

git diff --check origin/main...HEAD
git diff --check
node scripts/run-oxlint.mjs --tsconfig config/tsconfig/oxlint.core.json src/agents/pi-embedded-runner/run/runtime-context-prompt.ts src/agents/pi-embedded-runner/run/runtime-context-prompt.test.ts
pnpm check:changed

Alix-007 · 2026-05-26T11:44:38Z

This is a focused fix for #86849. It keeps before_prompt_build prompt context adjacent to the current user prompt in embedded runs, while preserving the existing hidden handling for marked OpenClaw internal runtime-context blocks.

Local verification:

git diff --check origin/main...HEAD
OPENCLAW_VITEST_MAX_WORKERS=1 node scripts/run-vitest.mjs src/agents/pi-embedded-runner/run/runtime-context-prompt.test.ts src/agents/pi-embedded-runner/run/attempt.spawn-workspace.context-engine.test.ts --pool forks --maxWorkers=1 --reporter=tap
pnpm check:changed

@clawsweeper re-review

clawsweeper · 2026-05-26T11:44:41Z

🦞👀
ClawSweeper picked this up.

Command router queued. I will update this comment with the next step.

Alix-007 · 2026-05-26T12:04:47Z

Follow-up pushed at 133f57660a with a regression covering the combined path: before_prompt_build prompt context stays visible in the model prompt while inter-session provenance and marked internal runtime context remain hidden in the runtime-context message.

Current local verification:

git diff --check origin/main...HEAD
OPENCLAW_VITEST_MAX_WORKERS=1 node scripts/run-vitest.mjs src/agents/pi-embedded-runner/run/runtime-context-prompt.test.ts src/agents/pi-embedded-runner/run/attempt.spawn-workspace.context-engine.test.ts --pool forks --maxWorkers=1 --reporter=tap
pnpm check:changed

@clawsweeper re-review

clawsweeper · 2026-05-26T12:04:50Z

🦞👀
ClawSweeper picked this up.

Command router queued. I will update this comment with the next step.

Alix-007 · 2026-05-26T14:48:41Z

Added structured real behavior proof to the PR body using the real exported runtime prompt splitter at head 133f57660a68.

The runtime output shows hook prepend/append text stays in the model prompt while the marked internal runtime context is excluded from the prompt and retained for hidden handling. No new commit was needed; code and prior targeted verification are unchanged.

@clawsweeper re-review

clawsweeper · 2026-05-26T14:48:45Z

🦞🧹
ClawSweeper re-review requested.

I asked ClawSweeper to review this item again.
Action: item re-review queued (workflow sweep.yml, event repository_dispatch).
Result: the existing ClawSweeper review comment will be edited in place when the review finishes.

Re-review progress:

State: Superseded
Detail: A newer re-review for this item started before this run finished, so GitHub cancelled this older run. Check the latest ClawSweeper run for the current result.
Run: https://github.com/openclaw/clawsweeper/actions/runs/26460916525
Updated: 2026-05-26T16:27:46.250Z

clawsweeper · 2026-05-26T14:51:35Z

Codex review: needs maintainer review before merge. Reviewed May 27, 2026, 6:58 PM ET / 22:58 UTC.

Summary
The PR separates embedded runner session/transcript prompts from model prompts so before_prompt_build prepend/append context remains model-visible while internal runtime-context blocks and transcript/UI text stay hidden.

PR surface: Source +360, Tests +500. Total +860 across 9 files.

Reproducibility: yes. source inspection of current main shows the splitter removes transcriptPrompt from effectivePrompt and treats the remainder as runtime context, which reproduces the reported hook-context stripping path when before_prompt_build changes the prompt.

Review metrics: none identified.

Merge readiness
Overall: 🐚 platinum hermit
Proof: 🦞 diamond lobster
Patch quality: 🐚 platinum hermit
Result: ready for maintainer review.

Overall follows the weaker of proof and patch quality, so missing proof can cap an otherwise strong patch.

Rank-up moves:

none

Risk before merge

Merging changes the embedded runner's provider-visible prompt boundary: existing hook/context-engine behavior can differ because the model prompt and persisted transcript prompt are now intentionally separate.
The patch is security-boundary sensitive because malformed or marked internal runtime-context text must never leak into the model prompt while hook context remains model-visible.

Maintainer options:

Accept the prompt-boundary change (recommended)
Maintainers can merge once they are comfortable that the current tests and proof cover the intended separation between transcript, model prompt, and hidden runtime context.
Ask for broader runtime proof
Maintainers can request an additional real embedded-runner scenario if they want proof beyond the exported helper and CI checks before accepting the boundary change.

Next step before merge
No ClawSweeper repair lane is needed; the remaining action is maintainer merge judgment for a compatibility- and security-sensitive prompt-boundary fix.

Security
Cleared: No concrete supply-chain or credential-handling problem was found; the prompt-hidden-context boundary is tracked as merge risk rather than a patch defect.

Review details

Best possible solution:

Land the current approach after maintainer acceptance of the transcript/model/runtime separation and current-head checks, keeping the focused regression coverage with the prompt-boundary change.

Do we have a high-confidence way to reproduce the issue?

Yes: source inspection of current main shows the splitter removes transcriptPrompt from effectivePrompt and treats the remainder as runtime context, which reproduces the reported hook-context stripping path when before_prompt_build changes the prompt.

Is this the best way to solve the issue?

Yes, with maintainer acceptance: separating session/transcript prompt text from model prompt text is the narrow maintainable fix, and the PR adds focused tests for hidden context extraction and leak prevention.

AGENTS.md: found and applied where relevant.

Codex review notes: model gpt-5.5, reasoning high; reviewed against 550a9b459a35.

Label changes

Label justifications:

P2: This is a normal-priority agent prompt/session-state bug fix with a bounded runtime surface but meaningful user impact.
merge-risk: 🚨 compatibility: The PR intentionally separates the model-visible prompt from the persisted transcript prompt, which can affect existing hook and context-engine behavior.
merge-risk: 🚨 security-boundary: The PR changes how marked internal runtime-context blocks are extracted, dropped, and prevented from reaching model-visible prompt text.
rating: 🐚 platinum hermit: Overall readiness is 🐚 platinum hermit; proof is 🦞 diamond lobster and patch quality is 🐚 platinum hermit.
status: 👀 ready for maintainer look: ClawSweeper has no concrete contributor-facing blocker left for this PR. Sufficient (terminal): The PR body provides terminal output from the real exported runtime prompt splitter showing hook context retained and valid, malformed, and multi-block internal runtime context kept out of the model prompt.
proof: sufficient: Contributor real behavior proof is sufficient. The PR body provides terminal output from the real exported runtime prompt splitter showing hook context retained and valid, malformed, and multi-block internal runtime context kept out of the model prompt.

Evidence reviewed

PR surface:

Source +360, Tests +500. Total +860 across 9 files.

View PR surface stats

Area	Files	Added	Removed	Net
Source	5	424	64	+360
Tests	4	526	26	+500
Docs	0	0	0	0
Config	0	0	0	0
Generated	0	0	0	0
Other	0	0	0	0
Total	9	950	90	+860

Acceptance criteria:

node scripts/run-vitest.mjs src/agents/embedded-agent-runner/run/runtime-context-prompt.test.ts src/agents/embedded-agent-runner/run/attempt.spawn-workspace.context-engine.test.ts --pool forks --maxWorkers=1 --reporter=tap
node scripts/run-oxlint.mjs --tsconfig config/tsconfig/oxlint.core.json src/agents/embedded-agent-runner/run/runtime-context-prompt.ts src/agents/embedded-agent-runner/run/runtime-context-prompt.test.ts
pnpm check:changed

What I checked:

Root policy read: Root AGENTS.md was read in full; its ClawSweeper policy requires whole prompt/security-boundary review and treats prompt/auth/session/config boundary changes as compatibility-sensitive. (AGENTS.md:18, 550a9b459a35)
Scoped policy read: The scoped agent runner guide favors focused helper tests plus nearest runner/context-engine proof for embedded runner prompt assembly changes. (src/agents/embedded-agent-runner/run/AGENTS.md:1, 550a9b459a35)
Current-main bug path: Current main's splitter returns prompt: transcriptPrompt and moves the effective-prompt remainder into runtimeContext, which explains why hook prepend/append context is stripped out when a transcript prompt is present. (src/agents/embedded-agent-runner/run/runtime-context-prompt.ts:63, 550a9b459a35)
PR implementation path: At PR head, resolveRuntimeContextPromptParts accepts a separate modelPrompt, extracts marked internal runtime context, and returns distinct prompt, modelPrompt, and runtimeContext values. (src/agents/embedded-agent-runner/run/runtime-context-prompt.ts:65, 616440a5ed9b)
PR call-site path: At PR head, the embedded runner passes the pre-hook prompt into the runtime-context split while separately using promptForModel for model-visible hook context and promptForSession for transcript/session submission. (src/agents/embedded-agent-runner/run/attempt.ts:4163, 616440a5ed9b)
Regression coverage: The PR adds helper and runner tests for hook context staying model-visible, multiple hidden blocks, inline marker mentions, malformed fail-closed blocks, and transcript marker stripping before provider assembly. (src/agents/embedded-agent-runner/run/runtime-context-prompt.test.ts:41, 616440a5ed9b)

Likely related people:

steipete: git blame ties the current resolveRuntimeContextPromptParts implementation on main to 15b1e99, and the live PR head commits are signed by the same account. (role: introduced current splitter and recent PR committer; confidence: high; commits: 15b1e99df310, fff4263eb35c, 616440a5ed9b; files: src/agents/embedded-agent-runner/run/runtime-context-prompt.ts, src/agents/embedded-agent-runner/run/attempt.ts, src/agents/embedded-agent-runner/tool-result-context-guard.ts)
mbelinky: Recent commit 7299c56 touched the embedded runner attempt.ts area for sub-agent cwd/workspace separation, making this a useful routing candidate for adjacent runner behavior. (role: recent adjacent contributor; confidence: medium; commits: 7299c5695317; files: src/agents/embedded-agent-runner/run/attempt.ts)
martingarramon: PR review comments called out multi-block extraction and delimiter scanning risks, then confirmed the fail-closed delimiter behavior was addressed. (role: reviewer; confidence: medium; files: src/agents/embedded-agent-runner/run/runtime-context-prompt.ts, src/agents/internal-runtime-context.ts)

What the crustacean ranks mean

🦀 challenger crab: rare, exceptional readiness with strong proof, clean implementation, and convincing validation.
🦞 diamond lobster: very strong readiness with only minor maintainer review expected.
🐚 platinum hermit: good normal PR, likely mergeable with ordinary maintainer review.
🦐 gold shrimp: useful signal, but proof or patch confidence is still limited.
🦪 silver shellfish: thin signal; proof, validation, or implementation needs work.
🧂 unranked krab: not merge-ready because proof is missing/unusable or there are serious correctness or safety concerns.
🌊 off-meta tidepool: rating does not apply to this item.

Shiny media proof means a screenshot, video, or linked artifact directly shows the changed behavior. Runtime, network, CSP, and security claims still need visible diagnostics.

How this review workflow works

ClawSweeper keeps one durable marker-backed review comment per issue or PR.
Re-runs edit this comment so the latest verdict, findings, and automation markers stay together instead of adding duplicate bot comments.
A fresh review can be triggered by eligible @clawsweeper re-review comments, exact-item GitHub events, scheduled/background review runs, or manual workflow dispatch.
PR/issue authors and users with repository write access can comment @clawsweeper re-review or @clawsweeper re-run on an open PR or issue to request a fresh review only.
Maintainers can also comment @clawsweeper review to request a fresh review only.
Fresh-review commands do not start repair, autofix, rebase, CI repair, or automerge.
Maintainer-only repair and merge flows require explicit commands such as @clawsweeper autofix, @clawsweeper automerge, @clawsweeper fix ci, or @clawsweeper address review.
Maintainers can comment @clawsweeper explain to ask for more context, or @clawsweeper stop to stop active automation.

clawsweeper · 2026-05-26T14:56:00Z

ClawSweeper PR egg

✨ Hatched: 🥚 common Moonlit Patch Peep

Hatch command

Comment @clawsweeper hatch when this PR is hatchable.

Hatchability rules:

Merged PRs are hatchable.
Open PRs are hatchable when they are status: 👀 ready for maintainer look, status: 🚀 automerge armed, or labeled clawsweeper:automerge.
Closed unmerged PRs are hatchable only when one of those hatchable labels is still present in the durable record.

Rarity: 🥚 common.
Trait: stacks clean commits.
Image traits: location branch lighthouse; accessory lint brush; palette violet, aqua, and starlight; mood focused; pose pointing at a small proof artifact; shell glossy opal shell; lighting cool dashboard glow; background tiny shells and proof notes.
Share on X: post this hatch
Copy: My PR egg hatched a 🥚 common Moonlit Patch Peep in ClawSweeper.

What is this egg doing here?

Eggs appear after the PR passes real-behavior proof. It is here for vibes, not verdicts: it does not change labels, ratings, merge decisions, or automation.
The shell reacts to review momentum: open follow-up work warms it up, re-review makes it wobble, and a clean final review lets it hatch.
Hatchability usually comes from sufficient real-behavior proof, no blocking P0/P1/P2 findings, no security attention needed, and clean correctness. A merged PR is already final, so merge makes the egg hatchable independently.
The hatch is seeded from this repository and PR number, so the same PR keeps the same creature; the reviewed head SHA can only change safe visual details.
Rarity is just collectible sparkle: 🥚 common, 🌱 uncommon, 💎 rare, ✨ glimmer, and 🌈 legendary.

martingarramon

Capturing effectivePrompt before annotateInterSessionPromptText runs ensures hook context reaches the model prompt regardless of annotation order.

extractMarkedRuntimeContext iterates over all BEGIN/END pairs in a for (;;) loop. Existing tests exercise single-block extraction ("keeps hidden runtime context separate from prompt-local additions"), but a prompt assembled with two separate INTERNAL_RUNTIME_CONTEXT_BEGIN/END blocks isn't covered. Add a multi-block test to confirm the remaining splicing is correct across iterations.

Edge case: findDelimitedRuntimeContextTokenIndex recurses each time the delimiter appears in text but not on its own line. An iterative version would remove the stack-growth risk for deeply nested or repeated delimiter occurrences.

clawsweeper · 2026-05-27T01:35:10Z

🦞🧹
ClawSweeper re-review requested.

I asked ClawSweeper to review this item again.
Action: item re-review queued (workflow sweep.yml, event repository_dispatch).
Result: the existing ClawSweeper review comment will be edited in place when the review finishes.

Re-review progress:

State: Complete
Detail: The targeted re-review finished, the durable review comment was updated, and the synced verdict was routed.
Run: https://github.com/openclaw/clawsweeper/actions/runs/26485432101
Updated: 2026-05-27T01:44:36.594Z

martingarramon

P1 fail-closed behavior confirmed: when extractMarkedRuntimeContext finds an unterminated <<<BEGIN_OPENCLAW_INTERNAL_CONTEXT>>> (no matching END), it returns prompt: before — the text before the marker — and the orphaned block content is dropped from the model prompt. The unterminated-block test locks this contract.

The iterative findDelimitedRuntimeContextTokenIndex preserves the line-delimited semantics of the recursive version. Multi-block extraction and inline-marker preservation both have regression coverage. LGTM.

martingarramon · 2026-05-27T13:06:48Z

P1 is resolved in 1d17a3f5596c. findDelimitedRuntimeContextTokenIndex anchors tokens to line boundaries — a token only matches when it starts after \n (or at position 0) and ends before \n/\r (or at EOF). The 250-inline-marker test confirms that mid-line occurrences don't match.

The fail-closed path (end === -1) returns { prompt: before.trimEnd() } — everything from the BEGIN marker onward is dropped. The "fails closed for unterminated" test locks the contract: an unterminated block returns { prompt: "visible ask" } with no runtimeContext.

modelPrompt carrying the pre-annotation prompt keeps prependContext/appendContext in the model prompt while inter-session provenance lands in runtimeContext. The new integration test pins both sides of the separation.

CI clean (86/86). LGTM.

clawsweeper · 2026-05-27T22:39:46Z

🦞🧹
ClawSweeper re-review requested.

I asked ClawSweeper to review this item again.
Action: item re-review queued (workflow sweep.yml, event repository_dispatch).
Result: the existing ClawSweeper review comment will be edited in place when the review finishes.

Re-review progress:

State: Superseded
Detail: A newer re-review for this item started before this run finished, so GitHub cancelled this older run. Check the latest ClawSweeper run for the current result.
Run: https://github.com/openclaw/clawsweeper/actions/runs/26542959926
Updated: 2026-05-27T22:46:07.698Z

clawsweeper · 2026-05-27T22:53:04Z

🦞🧹
ClawSweeper re-review requested.

I asked ClawSweeper to review this item again.
Action: item re-review queued (workflow sweep.yml, event repository_dispatch).
Result: the existing ClawSweeper review comment will be edited in place when the review finishes.

Re-review progress:

State: Complete
Detail: The targeted re-review finished, the durable review comment was updated, and the synced verdict was routed.
Run: https://github.com/openclaw/clawsweeper/actions/runs/26543485225
Updated: 2026-05-27T22:58:29.280Z

Co-authored-by: Alix-007 <267018309+Alix-007@users.noreply.github.com>

steipete · 2026-05-27T23:29:49Z

Landed via squash as 8b7a482. Thanks @Alix-007.

Proof:

rebased head 9715d3a
GitHub CI run 26544578760 passed
local tsgo/test-types, oxfmt, oxlint, git diff check passed
autoreview clean, no accepted/actionable findings

Fixes embedded agent prompt handling so before_prompt_build prepend/append context stays prompt-local: visible transcripts keep the user prompt, provider/model prompts keep hook context, and runtime/system context stays separate. Local verification: - git diff --check - fnm exec --using v22.22.2 pnpm exec oxfmt --check src/agents/embedded-agent-runner/tool-result-context-guard.ts src/agents/embedded-agent-runner/tool-result-context-guard.test.ts - fnm exec --using v22.22.2 node scripts/run-oxlint.mjs --tsconfig config/tsconfig/oxlint.core.json src/agents/embedded-agent-runner/tool-result-context-guard.ts src/agents/embedded-agent-runner/tool-result-context-guard.test.ts - fnm exec --using v22.22.2 pnpm tsgo:test:src - autoreview clean: no accepted/actionable findings CI verification: - GitHub CI run 26544578760 passed on rebased head 9715d3a Co-authored-by: Alix-007 <267018309+Alix-007@users.noreply.github.com>

…026.5.28) (#759) This PR contains the following updates: | Package | Update | Change | |---|---|---| | [ghcr.io/openclaw/openclaw](https://openclaw.ai) ([source](https://github.com/openclaw/openclaw)) | patch | `2026.5.27` → `2026.5.28` | --- ### Release Notes <details> <summary>openclaw/openclaw (ghcr.io/openclaw/openclaw)</summary> ### [`v2026.5.28`](https://github.com/openclaw/openclaw/blob/HEAD/CHANGELOG.md#2026528) [Compare Source](openclaw/openclaw@v2026.5.27...v2026.5.28) ##### Highlights - Agent and Codex runtime recovery is steadier: subagents keep cwd/workspace separation, hook context stays prompt-local, session locks release on timeout abort while live OpenClaw locks survive cleanup, stale restart continuations are avoided, and Codex app-server/helper failures no longer tear down shared runtime state. ([#87218](openclaw/openclaw#87218), [#86875](openclaw/openclaw#86875), [#87409](openclaw/openclaw#87409), [#87399](openclaw/openclaw#87399), [#87375](openclaw/openclaw#87375), [#88129](openclaw/openclaw#88129)) - Channel delivery and session identity got safer across outbound plugin hooks, Matrix room ids, iMessage reactions/approvals, Slack final replies, Discord recovered tool warnings, runtime-config message actions, WhatsApp profile auth roots, Telegram polling, and Microsoft Teams service URL trust checks. ([#73706](openclaw/openclaw#73706), [#75670](openclaw/openclaw#75670), [#87366](openclaw/openclaw#87366), [#87451](openclaw/openclaw#87451), [#87334](openclaw/openclaw#87334), [#84535](openclaw/openclaw#84535), [#82492](openclaw/openclaw#82492), [#83304](openclaw/openclaw#83304), [#87160](openclaw/openclaw#87160)) - Mobile and chat surfaces got a broader refresh: the iOS Pro UI, hosted push relay default, realtime Talk tab playback, Gateway chat transport, onboarding, Talk permissions, WebChat reconnect delivery, and session picker behavior now preserve more state across reconnects and empty searches. ([#87367](openclaw/openclaw#87367), [#87531](openclaw/openclaw#87531), [#87682](openclaw/openclaw#87682), [#88096](openclaw/openclaw#88096), [#88105](openclaw/openclaw#88105)) Thanks [@ngutman](https://github.com/ngutman) and [@BunsDev](https://github.com/BunsDev). - Browser, channel, and automation inputs are stricter: Browser tool timeouts, viewport/tab indices, Gateway ports, cron retry handling, Discord component ids, schema array refs, Telegram callback pages, and channel progress callbacks now reject malformed values earlier and preserve the intended delivery context. ([#82887](openclaw/openclaw#82887)) - Provider, media, and document coverage expands with Claude Opus 4.8, Fal Krea image schemas, NVIDIA featured models, MiniMax streaming music responses, encrypted PDF extraction, voice model catalogs, GitHub Copilot agent runtime support, and a Codex Supervisor plugin path for delegated Codex workflows. ([#87845](openclaw/openclaw#87845), [#87890](openclaw/openclaw#87890), [#80775](openclaw/openclaw#80775), [#84764](openclaw/openclaw#84764), [#87751](openclaw/openclaw#87751), [#87794](openclaw/openclaw#87794)) - CLI, auth, doctor, and provider paths fail faster and recover more clearly: malformed numeric/version options are rejected, workspace dotenv provider credentials are ignored, heartbeat defaults, OAuth/token lifetimes, and local service startup requests are bounded, agent auth health labels are clearer, legacy `api_key` auth profiles migrate to canonical form, and restart guidance is actionable. ([#87398](openclaw/openclaw#87398), [#86281](openclaw/openclaw#86281), [#87361](openclaw/openclaw#87361), [#88133](openclaw/openclaw#88133), [#83655](openclaw/openclaw#83655), [#87559](openclaw/openclaw#87559), [#88088](openclaw/openclaw#88088), [#85924](openclaw/openclaw#85924)) Thanks [@vincentkoc](https://github.com/vincentkoc) and [@giodl73-repo](https://github.com/giodl73-repo). - Plugin and Gateway hot paths do less repeated work while preserving cache correctness for install records, config JSON parsing, tool search catalogs, session stores, manifest model rows, auto-enabled plugin config, browser tokens, viewer assets, and release-split external plugin packages. ([#86699](openclaw/openclaw#86699)) - Release, QA, and E2E validation now bound more log, artifact, harness, and cross-OS waits so failing lanes produce proof instead of hanging or false-greening. ##### Changes - Status: show active subagent details in status output. - Diffs: split the default language pack and expand default Diffs language coverage while keeping the host floor aligned. ([#87370](openclaw/openclaw#87370), [#87372](openclaw/openclaw#87372)) Thanks [@RomneyDa](https://github.com/RomneyDa). - ClawHub: add plugin display names plus skill verification and trust surfaces. ([#87354](openclaw/openclaw#87354), [#86699](openclaw/openclaw#86699)) Thanks [@thewilloftheshadow](https://github.com/thewilloftheshadow) and [@Patrick-Erichsen](https://github.com/Patrick-Erichsen). - iOS: refresh the dev app with Pro Command, Chat, Agents, Settings, hosted push relay defaults, and realtime Talk playback wired to gateway sessions, diagnostics, chat, and realtime Talk. ([#87367](openclaw/openclaw#87367), [#88096](openclaw/openclaw#88096), [#88105](openclaw/openclaw#88105)) Thanks [@Solvely-Colin](https://github.com/Solvely-Colin) and [@ngutman](https://github.com/ngutman). - Docs: clarify Codex computer-use setup, paste-token stdin auth setup, macOS gateway sleep troubleshooting, native Codex hook relay recovery, container model auth, install deployment cards, device-token admin gating, CLI setup flow compatibility, Notte cloud browser CDP setup, and backport targets. ([#87313](openclaw/openclaw#87313), [#63050](openclaw/openclaw#63050), [#87685](openclaw/openclaw#87685)) Thanks [@bdjben](https://github.com/bdjben), [@liaoandi](https://github.com/liaoandi), and [@thewilloftheshadow](https://github.com/thewilloftheshadow). - PDF/tools: use ClawPDF for PDF extraction, support encrypted PDF extraction, and surface MCP structured content in agent tool results. ([#87670](openclaw/openclaw#87670), [#87751](openclaw/openclaw#87751)) - Providers: add Claude Opus 4.8 support, Fal Krea image model schemas, NVIDIA featured model catalogs, MiniMax streaming music responses, and provider-backed voice model catalogs. ([#87845](openclaw/openclaw#87845), [#87890](openclaw/openclaw#87890), [#80775](openclaw/openclaw#80775), [#84764](openclaw/openclaw#84764), [#87794](openclaw/openclaw#87794)) Thanks [@eleqtrizit](https://github.com/eleqtrizit) and [@vincentkoc](https://github.com/vincentkoc). - Codex/GitHub: add the GitHub Copilot agent runtime and the Codex Supervisor plugin package. - Plugins: externalize GitHub Copilot and Tokenjuice as official install-on-demand plugins with npm and ClawHub publish metadata. - Workboard: add agent coordination tools for tracking and handing off active agent work. - Discord: show commentary in progress drafts so live Discord runs expose useful in-progress context. ([#85200](openclaw/openclaw#85200)) - Plugin SDK: add a reply payload sending hook for plugins that need to deliver channel-owned replies and flatten package types for SDK declarations. ([#82823](openclaw/openclaw#82823), [#87165](openclaw/openclaw#87165)) Thanks [@piersonr](https://github.com/piersonr) and [@RomneyDa](https://github.com/RomneyDa). - Policy: add policy comparison, ingress-channel conformance, and sandbox-posture conformance checks. ([#85572](openclaw/openclaw#85572), [#85744](openclaw/openclaw#85744), [#86768](openclaw/openclaw#86768)) ##### Fixes - Agents: fall back to local config pruning when the optional `agents delete` Gateway probe cannot authenticate, so offline installs can still delete agents without removing shared workspaces. - Tighten phone-control mutation authorization \[AI]. ([#87150](openclaw/openclaw#87150)) Thanks [@pgondhi987](https://github.com/pgondhi987). - Clarify directive persistence authorization policy \[AI]. ([#86369](openclaw/openclaw#86369)) Thanks [@pgondhi987](https://github.com/pgondhi987). - Agents/Codex: keep spawned agent cwd/workspace state separated, forward ACP spawn attachments, keep hook context prompt-local, release session locks on timeout abort and runtime teardown without deleting live OpenClaw-owned locks during cleanup, avoid session event queue self-wait, clean up exec abort listeners, stream assistant deltas incrementally, recover raw missing-thread compaction failures, preserve rotated compaction session identity, keep compaction-timeout snapshots continuable, preserve shared app-server state across startup or helper failures, keep native hook relay alive across restarts and prune stale bridge files, close native hook relay replacement races, keep Claude live tool progress visible for watchdog recovery, suppress abandoned requester completion handoff, route workspace memory through tools, resolve Codex runtime models first, report quarantined dynamic tools, format `skills` command output, bind node auto-review to prepared plans, retry Claude CLI transcript probes, and bound compaction/steering retries. ([#87218](openclaw/openclaw#87218), [#86875](openclaw/openclaw#86875), [#86123](openclaw/openclaw#86123), [#88129](openclaw/openclaw#88129), [#87399](openclaw/openclaw#87399), [#87375](openclaw/openclaw#87375), [#72574](openclaw/openclaw#72574), [#87383](openclaw/openclaw#87383), [#87400](openclaw/openclaw#87400), [#83022](openclaw/openclaw#83022), [#87671](openclaw/openclaw#87671), [#87738](openclaw/openclaw#87738), [#87747](openclaw/openclaw#87747), [#87706](openclaw/openclaw#87706), [#87546](openclaw/openclaw#87546), [#87541](openclaw/openclaw#87541), [#81048](openclaw/openclaw#81048)) Thanks [@mbelinky](https://github.com/mbelinky), [@Alix-007](https://github.com/Alix-007), [@luoyanglang](https://github.com/luoyanglang), [@yetval](https://github.com/yetval), [@sjf](https://github.com/sjf), [@joshavant](https://github.com/joshavant), [@benjamin1492](https://github.com/benjamin1492), [@c19354837](https://github.com/c19354837), [@fuller-stack-dev](https://github.com/fuller-stack-dev), [@pfrederiksen](https://github.com/pfrederiksen), and [@dodge1218](https://github.com/dodge1218). - Codex Supervisor: keep real-home app-server MCP session listing on the loaded state path, bound stored history scans, and close WebSocket probes cleanly. - Channels: thread canonical session keys into outbound hooks, preserve Matrix room-id case, keep fallback tool warnings mention-inert, retain delivered Slack final replies during late cleanup, continue iMessage polling after denied reactions, suppress duplicate native exec approvals, resolve Gateway message actions against the active runtime config, preserve Telegram SecretRef prompt config and polling keepalives, preserve WhatsApp profile auth roots, QR display, document filenames, and plugin hook config, suppress Discord recovered tool warnings, preserve the Discord voice outbound helper, cap Discord/Signal/Zalo channel request and container timeouts, and block untrusted Teams service URLs while keeping TeamsSDK patterns aligned. ([#73706](openclaw/openclaw#73706), [#75670](openclaw/openclaw#75670), [#87366](openclaw/openclaw#87366), [#87451](openclaw/openclaw#87451), [#87465](openclaw/openclaw#87465), [#87334](openclaw/openclaw#87334), [#84535](openclaw/openclaw#84535), [#76262](openclaw/openclaw#76262), [#83304](openclaw/openclaw#83304), [#82492](openclaw/openclaw#82492), [#87581](openclaw/openclaw#87581), [#77114](openclaw/openclaw#77114), [#86426](openclaw/openclaw#86426), [#85529](openclaw/openclaw#85529), [#87160](openclaw/openclaw#87160)) Thanks [@zeroaltitude](https://github.com/zeroaltitude), [@lukeboyett](https://github.com/lukeboyett), [@jarvis-mns1](https://github.com/jarvis-mns1), [@xiaotian](https://github.com/xiaotian), [@funmerlin](https://github.com/funmerlin), [@joshavant](https://github.com/joshavant), [@eleqtrizit](https://github.com/eleqtrizit), [@heyitsaamir](https://github.com/heyitsaamir), [@amittell](https://github.com/amittell), [@lidge-jun](https://github.com/lidge-jun), [@liorb-mountapps](https://github.com/liorb-mountapps), [@masatohoshino](https://github.com/masatohoshino), [@bladin](https://github.com/bladin), and [@giodl73-repo](https://github.com/giodl73-repo). - CLI/auth/doctor/providers: reject malformed numeric/timeout/subcommand-version inputs, ignore workspace dotenv provider credentials, wait for respawn child shutdown, bound heartbeat defaults plus Codex, GitHub Copilot, OpenAI, Anthropic, Google, Feishu, LM Studio, MiniMax, Xiaomi TTS, and local-provider OAuth/token/model requests, harden Codex auth probes, label auth health by agent, preserve explicit agentRuntime pins during Codex model migration, warm provider auth off the main thread, honor Codex response timeouts, stop migrating current Claude Haiku 4.5 profiles to Sonnet, bound local service startup, resolve GPT-5.5 without cached catalog, migrate legacy memory auto-provider config, rewrite non-canonical `api_key` auth profiles, and make doctor restart follow-ups actionable. ([#87398](openclaw/openclaw#87398), [#86281](openclaw/openclaw#86281), [#87361](openclaw/openclaw#87361), [#88133](openclaw/openclaw#88133), [#83655](openclaw/openclaw#83655), [#87559](openclaw/openclaw#87559), [#87719](openclaw/openclaw#87719), [#88088](openclaw/openclaw#88088), [#85924](openclaw/openclaw#85924), [#84362](openclaw/openclaw#84362)) Thanks [@Patrick-Erichsen](https://github.com/Patrick-Erichsen), [@samzong](https://github.com/samzong), [@giodl73-repo](https://github.com/giodl73-repo), [@alkor2000](https://github.com/alkor2000), [@mmaps](https://github.com/mmaps), [@nxmxbbd](https://github.com/nxmxbbd), and [@vincentkoc](https://github.com/vincentkoc). - Gateway/security/session state: expire browser tokens after auth rotation, scope assistant idempotency dedupe, drain probe client closes, avoid stale restart continuation reuse, preserve retry-after fallbacks and stale rate-limit cooldown probes, bound webchat image and artifact transcript scans, include seconds in inbound metadata timestamps, clear completed session active runs, clear stale chat stream buffers, and evict current plugin-state namespaces at row caps. ([#87810](openclaw/openclaw#87810), [#87833](openclaw/openclaw#87833), [#75089](openclaw/openclaw#75089)) Thanks [@joshavant](https://github.com/joshavant) and [@litang9](https://github.com/litang9). - Config/parsing/network: reject partial numeric parsing, parse provider/Discord retry headers and dates strictly, honor IPv6 and bare IPv6 `no_proxy` entries, preserve empty plugin allowlists, canonicalize secret target array indexes, and reject malformed media content lengths, inspected TCP ports, marketplace content lengths, cron epochs, sandbox stat fields, unsafe duration values, empty config path segments, noncanonical schema array refs, unsafe Telegram callback pages, and invalid Teams attachment-fetch DNS targets. ([#87883](openclaw/openclaw#87883)) Thanks [@zhangguiping-xydt](https://github.com/zhangguiping-xydt). - Browser/input hardening: reject invalid tab indexes, excessive viewport resizes, explicit zero CDP ports, malformed geolocation options, unsafe screenshot or permission-grant timeouts, loose response-body limits, invalid cookie expiries, and non-finite Browser tool delays/timeouts. - Cron/automation: retry recurring jobs after transient model rate limits before waiting for the next scheduled slot, and preflight model fallbacks before skipping scheduled work. ([#82887](openclaw/openclaw#82887)) Thanks [@chen-zhang-cs-code](https://github.com/chen-zhang-cs-code). - Auto-reply/directives: respect provider and relayed channel metadata during directive persistence so channel-originated decisions keep their intended context. ([#87683](openclaw/openclaw#87683)) - WhatsApp: resolve the auth directory from the active profile so profile-scoped WhatsApp installs do not drift to the wrong credential root. ([#82492](openclaw/openclaw#82492)) Thanks [@lidge-jun](https://github.com/lidge-jun). - Gateway/session state: clear completed session active runs, avoid cold-loading providers for MCP inventory, cache single-session child indexes, cap handshake timers, and bound preauth, auth-guard, media, transcript, readiness, and port options. - Channels/replies: preserve channel-owned progress callbacks when verbose output is off, keep group-room progress suppression intact, prefer external session delivery context, escape Discord component id delimiters, force final TUI chat repaints, show Slack reasoning previews, and normalize Discord/Matrix/Mattermost channel numeric options. ([#87476](openclaw/openclaw#87476), [#87423](openclaw/openclaw#87423)) - Agents/tool args: harden smart-quoted argument repair for edit arrays and exact escaped arguments so model-produced tool calls recover without corrupting valid input. ([#86611](openclaw/openclaw#86611)) Thanks [@ferminquant](https://github.com/ferminquant). - Providers/agents: preserve seeded Anthropic signatures, preserve signed thinking payloads, concatenate signature-delta chunks, preserve DeepSeek `reasoning_content` replay across tier suffixes, apply OpenRouter strict9 ids to Mistral routes, promote Ollama plain-text tool calls, load NVIDIA featured model catalogs, stream MiniMax music generation responses, and recover empty preflight compaction. ([#87593](openclaw/openclaw#87593), [#87493](openclaw/openclaw#87493), [#80775](openclaw/openclaw#80775), [#84764](openclaw/openclaw#84764)) Thanks [@Pluviobyte](https://github.com/Pluviobyte) and [@eleqtrizit](https://github.com/eleqtrizit). - Media/images: skip CLI image cache refs when resolving generated images, allow trusted generated HTML attachments, and bound generated video downloads so stale refs and slow providers fail cleanly. ([#87523](openclaw/openclaw#87523), [#87982](openclaw/openclaw#87982)) - File transfer: handle late tar stdin pipe errors after archive validation or unpacking has already settled. - Performance: trust install-record caches between reloads, prefer native JSON parsing, reuse unchanged tool-search catalogs, reuse gateway session and plugin metadata paths, skip unchanged store serialization, patch single-entry session writes, add precomputed session patch writers, reduce store clone allocations, cache manifest model catalog rows and auto-enabled plugin config, avoid full session snapshots for entry reads, defer configured Slack full startup, prefer bundled plugin dist entries, and slim current metadata identity caches. ([#87760](openclaw/openclaw#87760)) - Docker/release/QA: package runtime workspace templates, stream cross-OS served artifacts, preserve sparse Crabbox run artifacts, isolate npm plugin installs per package, reject incompatible package plugin API installs, drop the leftover root Sharp dependency from package manifests after the Rastermill migration, bound OpenClaw instance logs, plugin gauntlet relay logs, MCP channel buffers, kitchen-sink scans, agent-turn assertions, QA-Lab credential broker calls, QA Matrix substrate requests, and release scenario logs, and keep release/google live guards current. ([#87647](openclaw/openclaw#87647), [#87477](openclaw/openclaw#87477)) Thanks [@rohitjavvadi](https://github.com/rohitjavvadi) and [@vincentkoc](https://github.com/vincentkoc). - Release/CI: bound manual git fetches, ClawHub verifier responses, ClawHub owner metadata, dependency-guard error bodies, Parallels limits, startup/test/memory budget parsing, and diffs viewer build warnings so release lanes fail with useful proof instead of hanging. ([#87839](openclaw/openclaw#87839)) </details> --- ### Configuration 📅 **Schedule**: Branch creation - At any time (no schedule defined), Automerge - At any time (no schedule defined). 🚦 **Automerge**: Disabled by config. Please merge this manually once you are satisfied. ♻ **Rebasing**: Whenever PR becomes conflicted, or you tick the rebase/retry checkbox. 🔕 **Ignore**: Close this PR and you won't be reminded about these updates again. --- - [ ] If you want to rebase/retry this PR, check this box --- This PR has been generated by [Renovate Bot](https://github.com/renovatebot/renovate).  Reviewed-on: https://git.erwanleboucher.dev/eleboucher/homelab/pulls/759

Fixes embedded agent prompt handling so before_prompt_build prepend/append context stays prompt-local: visible transcripts keep the user prompt, provider/model prompts keep hook context, and runtime/system context stays separate. Local verification: - git diff --check - fnm exec --using v22.22.2 pnpm exec oxfmt --check src/agents/embedded-agent-runner/tool-result-context-guard.ts src/agents/embedded-agent-runner/tool-result-context-guard.test.ts - fnm exec --using v22.22.2 node scripts/run-oxlint.mjs --tsconfig config/tsconfig/oxlint.core.json src/agents/embedded-agent-runner/tool-result-context-guard.ts src/agents/embedded-agent-runner/tool-result-context-guard.test.ts - fnm exec --using v22.22.2 pnpm tsgo:test:src - autoreview clean: no accepted/actionable findings CI verification: - GitHub CI run 26544578760 passed on rebased head 9715d3a Co-authored-by: Alix-007 <267018309+Alix-007@users.noreply.github.com>

openclaw-barnacle Bot added the agents Agent runtime and tooling label May 26, 2026

openclaw-barnacle Bot added size: S proof: supplied External PR includes structured after-fix real behavior proof. labels May 26, 2026

martingarramon reviewed May 26, 2026

View reviewed changes

martingarramon reviewed May 27, 2026

View reviewed changes

steipete self-assigned this May 27, 2026

steipete force-pushed the issue-86849-prompt-hook-context branch from 1d17a3f to fff4263 Compare May 27, 2026 22:30

openclaw-barnacle Bot added size: L and removed size: M labels May 27, 2026

openclaw-barnacle Bot added size: XL and removed size: L labels May 27, 2026

steipete and others added 2 commits May 28, 2026 00:21

fix(agents): keep hook context prompt-local

8189c2f

Co-authored-by: Alix-007 <267018309+Alix-007@users.noreply.github.com>

fix(agents): strip transcript marker from model context

9715d3a

steipete force-pushed the issue-86849-prompt-hook-context branch from 616440a to 9715d3a Compare May 27, 2026 23:22

openclaw-barnacle Bot removed the proof: sufficient ClawSweeper judged the real behavior proof convincing. label May 27, 2026

steipete merged commit 8b7a482 into openclaw:main May 27, 2026
101 of 102 checks passed

github-actions Bot mentioned this pull request May 28, 2026

📡 Upstream Digest — 2026-05-28 02:26 UTC curtismercier/openclaw-mods#961

Open

grifjef mentioned this pull request May 28, 2026

fix(agents): keep before_prompt_build hook prependContext in user prompt on empty transcript (#87163) #87424

Closed

15 tasks

clawsweeper Bot mentioned this pull request Jun 8, 2026

feat: add runtime self context config and tool #90101

Open

solodmd mentioned this pull request Jun 10, 2026

fix(agents): restore runtime context fallback for heartbeat prompts #91817

Closed

Uh oh!

Conversation

Alix-007 commented May 26, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Real behavior proof

Verification

Uh oh!

Alix-007 commented May 26, 2026

Uh oh!

clawsweeper Bot commented May 26, 2026

Uh oh!

Alix-007 commented May 26, 2026

Uh oh!

clawsweeper Bot commented May 26, 2026

Uh oh!

Alix-007 commented May 26, 2026

Uh oh!

clawsweeper Bot commented May 26, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

clawsweeper Bot commented May 26, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

clawsweeper Bot commented May 26, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Hatch command

Uh oh!

martingarramon left a comment

Choose a reason for hiding this comment

Uh oh!

clawsweeper Bot commented May 27, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

martingarramon left a comment

Choose a reason for hiding this comment

Uh oh!

martingarramon commented May 27, 2026

Uh oh!

clawsweeper Bot commented May 27, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

clawsweeper Bot commented May 27, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

steipete commented May 27, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Alix-007 commented May 26, 2026 •

edited

Loading

clawsweeper Bot commented May 26, 2026 •

edited

Loading

clawsweeper Bot commented May 26, 2026 •

edited

Loading

clawsweeper Bot commented May 26, 2026 •

edited

Loading

clawsweeper Bot commented May 27, 2026 •

edited

Loading

clawsweeper Bot commented May 27, 2026 •

edited

Loading

clawsweeper Bot commented May 27, 2026 •

edited

Loading