agentleFS
curv-institutecurv-institute/atlas/.claude/skills/council/SKILL.md

council

Skill0 starsChanged 4 months agoMITView on GitHub

Convene an adversarial review (ideally multi-agent) to verify a result, vet a design, or make a go/no-go call - refute don't confirm, re-derive load-bearing numbers independently, hunt the known confounds, and scrutinize any positive hardest of all. Use when the user says "/council", "adversarially review/verify this", "is this real / break this claim", "vet this plan", or "should we go/no-go on X". Operationalizes OPERATING_GUIDE.md §1.6 / §3 / §6.

167 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use agent-payment-x402, defi-amm-security and 10 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • TironFoxSkill

    agent-payment-x402

    Add x402 payment execution to AI agents — per-task budgets, spending controls, and non-custodial wallets via MCP tools. Use when agents need to pay for APIs, services, or other agents.

    Also in 5 projects that use this

  • RojiosSkill

    defi-amm-security

    Security checklist for Solidity AMM contracts, liquidity pools, and swap flows. Covers reentrancy, CEI ordering, donation or inflation attacks, oracle manipulation, slippage, admin controls, and integer math.

    Also in 6 projects that use this

  • cogine-aiSkill

    modernize

    For getting old prs up to date. Use when asked to rebase a pr or "modernize".

    Also in 9 projects that use this

  • cogine-aiSkill

    leaving-pr-comments

    Use whenever leaving comments on GitHub.

    Also in 8 projects that use this

  • automagik-devSkill

    genie-hacks

    Browse, search, and contribute community hacks — real-world patterns for provider switching, teams, skills, hooks, cost optimization, and more.

    Also in 3 projects that use this

  • automagik-devSkill

    workfly

    Discover a procedure and build its saved workflow — dynamic discovery, drafted script, adversarial verification, landed in the catalog.

    Also in 3 projects that use this

  • +1 more from cogine-ai

The file

We split this skill into the rules it gives an agent

1 rule. 0 are common, 1 is distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 150 edited versions we read keep these.

  • 1

    Distinctive

    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
  1. Distinctive:

    Pre-commit the verdict rule: before looking (what evidence = PASS / FAIL / MARGINAL).

    line 23

“Common” is measured: kept by at least 6 in 10 of the 150 edited versions of this file we could read. “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

/council - adversarial review, verify, or go/no-go

Try to break a claim, design, or decision before it becomes a claim. The bias is refutation: a finding that survives a real attempt to kill it is worth far more than one that was confirmed. Run as a panel of independent, refutation-minded reviewers (use parallel sub-agents if available).

Procedure

  1. State the claim precisely - the exact quantity, scope, N, units, and the artifact it rests on.
  2. Assign independent angles (one reviewer each; don't let them confirm each other):
    • Re-derivation - recompute every load-bearing number from the raw artifact by an independent route. Trust no number that came only from a script or a sub-agent.
    • Confound hunt - walk the OPERATING_GUIDE.md §3 catalog: currency/units, convergence/ resolution, same-corpus/same-split, capacity mismatch, empty-output/cwd. A verdict that flips on a unit conversion is an automatic red flag.
    • Instrument check - can the instrument even resolve the effect? Power / dynamic range / CI vs signal. Pre-stated thresholds, no fail-to-pass tuning?
    • Steelman the null - try hardest to make the positive go away.
  3. Pre-commit the verdict rule before looking (what evidence = PASS / FAIL / MARGINAL).
  4. Adjudicate - PASS only if it survives all angles. Return the verdict + the surviving evidence + any corrections, and update the §6 checklist.

Output

A written verdict: claim, attempts to refute, what survived, what was corrected/withdrawn, and the final PASS / FAIL / MARGINAL with scope. Record withdrawals additively in CURRENT_STATE.md.

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.