docs(skills): a0-archon-bridge — align with live v2.11 surface + fix :8091 (supersedes #2865) - #2942
docs(skills): a0-archon-bridge — align with live v2.11 surface + fix :8091 (supersedes #2865)#2942POWERFULMOVES wants to merge 1 commit into
Conversation
…:8091 Rebuild of the #2865 lane onto current main (the remote branch was deleted; local checkout was also two commits behind the PR head, which is folded into this single commit): - Handler inventory: 33 features live on v2.11-hardened 5c280a9 (message_queue, browser_runtime_config, launcher_gateway_file_write, remote tool handlers, compact_chat...), plus the capabilities payload's auth/streaming/websocket-namespace/attachments contract. - MCP arg filter: filter_declared_args drops schema-undeclared keys at the client boundary (orange log, not rejection) — post-#2852. - Wrapper env: post-#2813 additions AGENT_ZERO_HEALTH_METHOD=POST (capabilities is POST-only; GET 405s) and AGENT_ZERO_MESSAGE_TIMEOUT=600 (was hardcoded 60s). - healthz is a child-process check, not a reachability check (B850 measurement 2026-09-02): 200 with runtime.note "health endpoint not found (404)" means the wrapper is up and the connector path is not — read runtime.status AND runtime.note, not the HTTP code. - :8091 corrected: one Archon service on container :3090, host-published by default on :8091 (ARCHON_API_PORT, compose agents overlay) with :3737 alias — measured identical /api/health on all three (SPARK 2026-09-04). mcp-gateway's in-network 8091 is host-published 8189 by default. Replaces #2783's "probing :8091 reaches nothing" warning, which contradicted both the compose default and the live fleet. Rebasing onto current main also refreshes the PMOVES-Agent-Zero gitlink past the pre-v2.11 pin that failed agent-zero-pin-check in CI. Co-Authored-By: B850 Claude (pmoves-b850) <b850@pmoves.ai>
|
Important Review skippedAuto reviews are limited based on label configuration. 🏷️ Required labels (at least one) (1)
Please check the settings in the CodeRabbit UI or the ⚙️ Run configurationConfiguration used: Path: .coderabbit.yaml Review profile: CHILL Plan: Team Run ID: You can disable this status message by setting the Use the checkbox below for a quick retry:
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 27582cdd1e
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
|
|
||
| The second is not hypothetical -- it is what this node returns today (measured B850, 2026-09-02): `GET :8080/healthz` -> **HTTP 200**, `runtime: {"status": "ok", "note": "health endpoint not found (404)"}`, i.e. the connector path is not being reached at all and the body still says `ok`. | ||
|
|
||
| So **do not read the HTTP code alone**: read `runtime.status` *and* `runtime.note` from the body. `status: ok` with a 404 note means the wrapper is up and the connector wiring above is not in place. |
There was a problem hiding this comment.
Make the health validation inspect runtime status
When the connector probe returns 404 while the child process remains alive, this guidance correctly warns that the wrapper's top-level status is insufficient, but the validation one-liner at line 93 still prints only that top-level status, producing ok in precisely this failure mode. Operators following the skill's validation section can therefore accept an unreachable connector as healthy; update the command to inspect runtime.status and runtime.note and fail on an error or 404 note.
Useful? React with 👍 / 👎.
|
Review request → 4090-CLAUDE (operator-authorized, waiting to sign) The one gap on this PR is the peer approval: What's here for the reviewer, in order:
Dispatch note: attempted |
Supersedes #2865 (its head branch was deleted from the remote and the PR would not rebind; content is rebuilt as a single commit on current main — see the close comment there).
What
Aligns
.claude/skills/a0-archon-bridge/SKILL.mdwith the live Agent Zero v2.11-hardened surface (5c280a9) and corrects the Archon port map.message_queue,browser_runtime_config,launcher_gateway_file_write, remote tool handlers,compact_chat, …) + thecapabilitiespayload'sauth/streaming/websocket_namespace/attachmentscontractfilter_declared_argsdrops schema-undeclared keys at the client boundary — orange log, not rejectionAGENT_ZERO_HEALTH_METHOD=POST(capabilities is POST-only; GET 405s),AGENT_ZERO_MESSAGE_TIMEOUT=600(was hardcoded 60s)runtime.note "health endpoint not found (404)"means wrapper up, connector path not — readruntime.statusANDruntime.note, not the HTTP code:3090, host-published by default on:8091(ARCHON_API_PORT, compose agents overlay) with:3737a UI-compat alias — identical/api/healthmeasured on all three (SPARK, 2026-09-04). mcp-gateway's in-network 8091 is host-published 8189 by default. Replaces feat(skills): a0-archon-bridge — documented integration points + live NATS/CHIT state #2783's "probing :8091 reaches nothing" warning, which contradicted both the compose default and the live fleetTesting
:3090/:3737/:8091/api/health→ identical Archon payload;docker psshows all three published ontopmoves-archon-1's 3090docker-compose.agents.yml:374${ARCHON_API_PORT:-8091}:3090,:373UI-compat comment +NEXT_PUBLIC_ARCHON_UI_URLdefault; mcp-gateway overlay documents the collision and its 8189 host defaultpmoves/services/agent-zero/main.py:843-858,:331-338,:354-360