agentleFS
iwe-orgiwe-org/dev-workspace/.claude/skills/verify/SKILL.md

verify

Skill11 starsChanged 2 days agoMITView on GitHub

Verify that implementation matches the graph — check a plan's tasks, spec requirements, and scenarios against the actual code before shipping, or sweep the whole workspace for doc↔code drift. Use when the user says "verify <plan>", "is this ready to ship", "audit the docs", or "check for drift".

189 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use instinct-export, instinct-import 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.
  • TheKREATORSSkill

    instinct-export

    Export instincts from project/global scope to a file

    Also in 3 projects that use this

  • TheKREATORSSkill

    instinct-import

    Import instincts from file or URL into project/global scope

    Also in 3 projects that use this

  • AviadPSkill

    risk-assess

    Assess the risk and blast radius of a proposed code change before implementing it. Evaluates CI impact, framework-wide vs localized scope, rollback path, and identifies blindspots. Use when user says "what's the risk", "is this safe", "assess the impact", "risk assess", "blast radius", or before applying framework-wide or CI-affecting changes. Do NOT use for code quality review (use code-quality) or plan validation (use design-validator).

    Also in 3 projects that use this

  • diego-tobalinaSkill

    build-fix

    Fix build and compilation errors systematically. Run when build fails.

    Also in 6 projects that use this

  • AaravChadhaSkill

    secure

    Security review with a confidence gate - a finding exists only with a concrete exploit scenario and a high/medium/low confidence rating; everything else goes to a worth-hardening list, never inflated. Sweeps five surfaces: auth gates, secrets hygiene, injection and unsafe sinks, unsafe deserialization/crypto/transport, and LLM tool-use trust boundaries. Reports only, never fixes. Use when the user asks for a security review or to check vulnerabilities, secrets, or auth.

    Also in 3 projects that use this

  • ArvorCoSkill

    specify

    Create feature specification from natural language description. Use when starting a new feature or creating a spec. Triggers on: create spec, specify feature, new feature spec.

    Also in 10 projects that use this

  • +1 more from TheKREATORS

The file

We split this skill into the rules it gives an agent

4 rules. 0 are common, 2 are distinctive, and 2 are very specific to this project.

  • 0

    Common

    Most of the 166 edited versions we read keep these.

  • 2

    Distinctive

    This author's own choices. Worth a look.

  • 2

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    No plan ("audit the docs", "check for drift") → audit mode (step 5).

    line 16
  2. Very specific:

    Select the plan: If ambiguous, list the plans under ## Active in data/plans.md (a just-finished plan may also sit under ## Done) and ask.

    line 17
  3. Very specific:

    Every claim cites path:line evidence or is explicitly labeled "unverified" — a requirement you couldn't trace is unverified, not failed, …

    line 68
  4. Distinctive:

    Run only the verification commands the plan itself names (tests, builds).

    line 71

“Common” is measured: kept by at least 6 in 10 of the 166 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

Verify implementation against the graph

The graph makes claims — tasks checked, requirements SHALL-ed, features marked implemented. This skill tests those claims against the code and reports what doesn't hold. It never fixes anything: the report is the deliverable, and the fixes belong to the skills and sessions it points at.

Steps

  1. Pick the mode. A plan named or inferable from conversation → plan mode. No plan ("audit the docs", "check for drift") → audit mode (step 5).
  2. Select the plan. If ambiguous, list the plans under ## Active in data/plans.md (a just-finished plan may also sit under ## Done) and ask. Announce: "Verifying: <plan>". Read the full plan, every spec named in its ## Spec changes, and the linked feature/bug docs (iwe retrieve -k data/plans/<key>).
  3. Check three dimensions, collecting issues as CRITICAL / WARNING / SUGGESTION, each with a concrete recommendation:
    • Completeness — every unchecked - [ ] task is a CRITICAL ("complete it, or tick it if already done"). Every spec in ## Spec changes exists and reflects the change (a back-ticked not-yet-created spec still pending at verify time is a CRITICAL). Every ### Requirement: in the touched specs has implementation evidence in the codebase — search for it; none found is a CRITICAL.
    • Correctness — map each requirement to path:line evidence and judge whether the implementation matches its SHALL statement (divergence is a WARNING with the file and lines to review). For each #### Scenario:, check the condition is handled in code and covered by a test (uncovered scenario is a WARNING). Run the plan's ## Verification commands and report their actual output — a failing command is a CRITICAL.
    • Coherence — the implementation follows the plan's ## Approach (a different approach that works is a WARNING: either the code or the plan should change); ## Out of scope items stayed out; the ## Authoring rules in data/product.md were honored.
  4. Report and stop. Issues ranked most severe first, then the verdict: ready to ship (zero CRITICAL — hand off to the ship skill) or the blocker list. Do not fix, tick, or edit anything.
  5. Audit mode — the same discipline over the whole graph, against the codebase:
    • Specs whose requirements the code now contradicts (sample the highest-traffic specs first: iwe find --references <key> counts).
    • Features marked implemented with no trace in the code, or whose tests no longer exist; shipped behavior with no feature doc.
    • Open bugs (no stage) — still reproducible? Point at ones whose ## Key references no longer exist. Recurring bugs (two or more affected_versions) get a WARNING asking why the first fix did not hold.
    • Learnings the code now contradicts — a correction is due.
    • Consistency: closed items still on an open list — iwe find --filter '{type: plan, stage: {$in: [done, cancelled]}}' --included-by data/plans -f keys and the same for {type: bug, stage: {$in: [done, cancelled]}} under data/bugs must print nothing; plans whose body claims they shipped (every box ticked) but carry no stage; iwe schema validate violations; a ## Broken Links section in iwe stats, or a data/ document under its ## Orphans section (iwe stats prints either section only when it has something to list).
    • Report in the same CRITICAL/WARNING/SUGGESTION format, grouped by fix owner: "run the ship skill on X", "update spec Y", "close bug Z".

Rules

  • Report, never fix — even a one-character hub-section fix is someone else's commit, so the audit trail stays clean.
  • Every claim cites path:line evidence or is explicitly labeled "unverified" — a requirement you couldn't trace is unverified, not failed, and says so.
  • Run only the verification commands the plan itself names (tests, builds); anything with side effects beyond the working tree needs the user's go-ahead.
  • Ship follows a clean verify; a verify with CRITICALs ends with the blocker list, not a softened verdict.

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.