agentleFS
youcislayoucisla/Agent-Foundry/skills/core/re-verify-findings/SKILL.md

re-verify-findings

Skill1 starsChanged 23 days agoMITView on GitHub

Before executing any audit finding, claimed bug, or reported issue, re-verify

The file

We split this skill into the rules it gives an agent

8 rules. About 0 are common, 8 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈8

    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
  1. Distinctive:

    Never execute a finding without re-verifying it against current code.

    line 16
  2. Distinctive:

    Before executing any plan, fix, or audit item

    line 23
  3. Distinctive:

    When user says "the audit found X" or "as reported by Y"

    line 24
  4. Distinctive:

    When inheriting work from another agent or model

    line 25
  5. Distinctive:

    Before deploying a fix for a claimed problem

    line 26
  6. Distinctive:

    Every claim from an external source gets a status before execution:

    line 32
  7. Distinctive:

    Do NOT accept the brief's severity without verifying.

    line 37
  8. Distinctive:

    When a claim is wrong or stale, state clearly:

    line 52

≈ “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

Re-Verify Findings

Never execute a finding without re-verifying it against current code.

When to Use

Activates on any prompt containing "audit", "fix", "bug", "issue", "according to", "was reported", "the plan says", "as stated in".

  • Before executing any plan, fix, or audit item
  • When user says "the audit found X" or "as reported by Y"
  • When inheriting work from another agent or model
  • Before deploying a fix for a claimed problem

Procedure

1 — Tag every claim as hypothesis

Every claim from an external source gets a status before execution:

  • ✅ confirmed (verified against live source)
  • 🔧 already fixed (verified current code is correct)
  • ❌ wrong (brief contradicts reality)

Do NOT accept the brief's severity without verifying.

2 — Verify against every live source available

SourceWhat to check
Repo (read_file/grep/git)Is the claimed code still present? Exact line number?
Git logWas it already fixed? When?
Live DB (SQL query)Schema, data, migration state match claim?
Live API (curl/browser)Does the deployed endpoint behave as claimed?
Stripe/3rd-party MCPDoes the connected account have what the claim asserts?
Browser (live app)Does the UI actually look/behave as described?

3 — Document discrepancies

When a claim is wrong or stale, state clearly:

Claim: [from the brief/audit]
Reality: [what you verified]
Implication: [plan change or skip]

This goes in the final report so the user understands why an item was skipped or modified.

Anti-patterns

Accepting the brief's claim at face value, executing, then discovering the fix was unnecessary. Wastes work and erodes trust.

  • Executing the finding exactly as written.
  • Accepting the audit's severity without checking.
  • Skipping a stale item without recording why.

Composes With

  • verify-first — broader discipline this is a subset of
  • pushback-when-wrong — the delivery vehicle for the discrepancy

Verification Checklist

  • Every finding has confirmed, fixed, wrong, or new status.
  • The claimed line or behavior was checked in current source.
  • Discrepancies state claim, reality, and implication.
  • Skipped items have a recorded reason.
  • Severity was verified, not inherited.

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.