peer-consult
Skill2 starsChanged 21 days agoMITView on GitHub
Delegate an explicitly requested question or review to a configured Claude or Codex CLI and bring the answer back into this conversation. Use for explicit requests to ask Claude or ask Codex to review. Also handles explicit developer-then-reviewer delegation using a configured role workflow. Do not trigger for ordinary local work, feature explanations, or inside a delegated Bridge worker.
1 other project uses a version of this.
Adopted by
Most recently taken up by
- dstotijn/agent-skillstheir own version, first seen Sep 21, 2026
Other versions1 project uses a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 1
- dstotijn/agent-skills★ 1same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
The file
We split this skill into the rules it gives an agent
2 rules. About 0 are common, 2 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- could be automatic: a setting could enforce it
- Distinctive: line 12
Use exactly the peer the user named.
- Distinctive: line 18
If files changed, inspect the manifest and relevant tests.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Peer consultation
Use this only as the interactive main agent in the project root containing both ./bridge.py and ./bridge.json. A delegated task copy has neither; do not search its parents for a launcher or re-delegate from it. The user's request authorizes only the explicitly requested consultations or roles, not recursive calls, extra targets, permission changes or original-file promotion.
For an explicit developer-and-reviewer request, read references/roles.md and use that sequence. For a single consultation, read references/usage.md. Execute the workflow yourself; do not tell the user to write a request document or copy commands for routine consultation.
- Use exactly the peer the user named. Ask only if the target or question cannot be inferred. Keep the configured endpoint, model, authentication and permission settings.
- Prepare only the question, relevant current facts/decisions, requested output language and exact project-relative source paths. Explain whether the request is advice/review or an authorized edit. For advice/review explicitly prohibit changes. Do not forward the entire conversation, raw logs, secrets or obsolete handoffs. The peer does not automatically know this conversation.
- Confirm that the necessary saved files are available under include (and any selection plan). Do not silently expand include/writable or tool permissions. If context is missing, explain what cannot be provided under the current configuration.
- Reuse the existing healthy runner. If it is stopped, stale, paused, or at its call limit, report the state and retain any submitted ID; do not bypass controls or launch duplicate runners. Runner control changes require task/session authorization.
- Submit once with a unique ID and use wait/result on that ID. A timeout is not permission to resubmit. Failures require inspection and a newly authorized request with a new ID.
- Bring the peer's actual answer back into the current conversation, identify which peer answered, and separate its findings from your assessment. Preserve requested language. If wait/result reports ERROR, TIMEOUT, INTERRUPTED, or an unavailable/partial answer, stop and report consultation failure. Do not mine raw logs for a substitute successful review or present your own analysis as the peer answer. DONE is not proof of correctness. Preserve the reviewer's caveats and unverified assumptions, including observations labeled non-blocking. If a caveat contradicts the requested contract, keep it open rather than summarizing the result as fully correct.
- If files changed, inspect the manifest and relevant tests. A consultation does not authorize apply; use only separately authorized, reviewed original paths. Do not automatically merge different peers' changes.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

