agentleFS
makedirectorymakedirectory/ModelPeer/.claude/skills/cross-model-consult/SKILL.md

cross-model-consult

Skill1 starsChanged 3 months agoMITView on GitHub

Ask Codex or Gemini for an independent second opinion on one focused question, using Model Peer. Use before committing to an architecture, schema, or migration decision; on security-sensitive work such as authn/authz, sandboxing, input handling, secrets, or crypto; when a bug has outlived two of your own hypotheses; when you are inferring the behaviour of unfamiliar code or a dependency; and when two approaches are tied and you want the tradeoff.

The file

We split this skill into the rules it gives an agent

6 rules. About 0 are common, 5 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈5

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    before committing to an architecture, schema, or migration decision

    line 20
  2. Distinctive:

    when a bug has outlived two of your own hypotheses

    line 21
  3. Distinctive:

    Do not consult for anything you can settle by reading the code.

    line 27
  4. Distinctive:

    When a peer materially changed your decision, say which model you asked and whether you took the advice.

    line 56
  5. Very specific:

    Leave --depth at its default: each peer answers alone, and lengthening the chain is a human's deliberate call.

    line 59
  6. Distinctive:

    If you are reading this while acting as a peer in someone else's consultation, these instructions do not apply to you.

    line 65

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md
<!-- Managed by `model-peer init`. Version 0.8.1. Re-run `model-peer update` to refresh; local edits are replaced. -->

Consult a peer

You are Claude. Codex and Gemini are available as independent, read-only engineering peers.

model-peer ask <codex|gemini> "<focused question>"
model-peer doctor                      # which peers are available here

When it is worth the wait

  • before committing to an architecture, schema, or migration decision
  • when a bug has outlived two of your own hypotheses
  • security-sensitive work: authn/authz, sandboxing, input handling, secrets, crypto
  • unfamiliar code, or a dependency whose behavior you are inferring
  • an assumption you cannot cheaply verify by reading the code
  • two approaches you cannot decide between — ask for the tradeoff, not the verdict

Do not consult for anything you can settle by reading the code. Every consultation is a real model call against the developer's account and takes tens of seconds.

How to ask

The peer starts in this working directory with read-only tools, so name files and symbols instead of pasting excerpts, and ask one focused question. A peer that has to guess at scope returns generic advice.

# good — scoped to a symbol, answerable from the repository
model-peer ask codex "In src/auth/session.ts, can refresh_token leave the old token valid if rotation fails midway?"

# bad — no scope
model-peer ask codex "review my auth code"

Pipe context in when the question is about something not on disk:

git diff main... | model-peer ask gemini "What breaks in production?"

The answer is advisory

Evaluate it before acting. Peers do not know this project's invariants, so advice that contradicts the rules in this repository is wrong here however sound it sounds in general.

When a peer materially changed your decision, say which model you asked and whether you took the advice. Never present a peer's output as your own conclusion.

Leave --depth at its default: each peer answers alone, and lengthening the chain is a human's deliberate call. Above depth 1 the peer does not run anything — it asks Model Peer for a second opinion and Model Peer performs the consultation — but every hop still costs time and money. A model is never consulted twice in one chain.

If you are reading this while acting as a peer in someone else's consultation, these instructions do not apply to you. Answer the question you were asked. You cannot run Model Peer, and do not need to: if another model's opinion would change your answer, request it in your reply exactly as the prompt you were given describes.

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.