agentleFS
aaronreneaaronrene/knowtation/.cursor/skills/deploy-verification-review/SKILL.md

deploy-verification-review

Skill1 starsChanged 2 months agoMITView on GitHub

Use before marking a phase DONE in {{docs.roadmap_path}} / regenerating HANDOVER when the session claims a live deploy, production (or staging-as-production) rollout, a public URL is "up" / "healthy" / "verified", or "shipped" meaning running outside the local repo (not merely code merged to a feature branch). Invoke: /deploy-verification-review. This is the live-deploy sibling of /build-verification-review. The kit still never deploys and never HTTP-probes production — it only records and optionally gates operator-supplied deploy/health claims (Mode C / require_deploy_health). Catch…

The file

We split this skill into the rules it gives an agent

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

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈0

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 6

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Use before marking a phase DONE in {{docs.roadmap_path}} / regenerating HANDOVER when the session claims a live deploy, production (or …

    line 3
  2. Very specific:

    When the build session claims deploy/health, BV V8 already requires a deploy_health artifact in the evidence entry — unchanged.

    line 65
  3. Very specific:

    When honesty.enabled: false, ledger append is skipped.

    line 72
  4. Very specific:

    Do not mark ship/DONE under require until pass

    line 91
  5. Very specific:

    No merge to {{vcs.git.main_branch}} on findings or blocked

    line 123
  6. Very specific:

    Do not waive /build-verification-review for Auto DONE

    line 125

≈ “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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Deploy verification review skill

Use before marking a phase DONE in {{docs.roadmap_path}} / regenerating HANDOVER when the session claims a live deploy, production (or staging-as-production) rollout, a public URL is "up" / "healthy" / "verified", or "shipped" meaning running outside the local repo (not merely code merged to a feature branch). Invoke: /deploy-verification-review.

This is the live-deploy sibling of /build-verification-review. The kit still never deploys and never HTTP-probes production — it only records and optionally gates operator-supplied deploy/health claims (Mode C / require_deploy_health).

Purpose

Catch ship/live-health dishonesty: "we shipped it" / "production is healthy" without a durable, content-addressed deploy_health record on the honesty ledger (when honesty is enabled).

GateWhenTool
Build verificationEvery Auto build DONE/build-verification-review (mandatory; never waived)
Deploy verificationShip / live-health claim presentThis skill
Mode C CI gateOpt-in require_deploy_health: requireok honesty-status --deploy-health PHASE_ID

When to use

Invoke when any of these claims appear in the deliverable, handover, or ROADMAP row:

  • live deploy / production (or staging-as-production) rollout
  • public or customer-facing URL is "up" / "healthy" / "verified"
  • "shipped" meaning running outside the local repo

Non-triggers: ordinary Auto builds that only claim code + tests + local CLI behavior (kit governance phases, Track O ceremony without a live consumer probe) use /build-verification-review alone. Track O ok upgrade-regime is a regime ceremony, not a P-deploy trigger.

Model

Always thinking-high — independent reviewer; preferably not the same session that claimed the ship. Never Auto for this gate.

Inputs (read all)

  1. Frozen spec — must authorize a ship/live-health claim (D1)
  2. Operator-supplied health-record bytes (or CI artifact) the verifier inspected
  3. {{docs.roadmap_path}} / {{docs.handover_path}} — ship wording vs Mode C
  4. Ledger (when honesty.enabled) — verification_evidence with ≥1 deploy_health artifact
  5. Git / Muse diff — no kit-side deploy or production HTTP probe helpers

Checklist (every item needs evidence)

#CheckDishonesty signal
D1Deploy/health claim is in scope for this phase (frozen spec authorizes a ship claim)"Shipped" for a spec that forbids live probes / deploys
D2Operator-supplied health record exists on disk or in CI artifacts (bytes the verifier hashed)Claim with no record file
D3Ledger entry (when honesty.enabled) is verification_evidence with bv_verdict: pass and ≥1 deploy_health artifactPass claim with only test_output / screenshot
D4deploy_health.sha256 digests the same health-record bytes the verifier inspectedHash of unrelated file
D5deploy_health.ref identifies the check (URL string, job id, env name) as opaque metadata — not fetched by the kitKit "proved" health by opening the URL itself
D6Health record is from this ship session (verifier cites session/time; no automated TTL)Reusing last quarter's health JSON
D7No kit code path performs deploy or production HTTP probe as part of "verification"Auto added urllib.request.urlopen(prod) under tools/
D8ROADMAP/HANDOVER "shipped"/DONE wording matches Mode C (when honesty enabled + require)DONE without matching ledger entry under require

Relationship to build-verification

  1. /build-verification-review remains mandatory for every Auto build DONE.
  2. When the build session claims deploy/health, BV V8 already requires a deploy_health artifact in the evidence entry — unchanged.
  3. This skill is the sibling gate for shipped/live claims. It does not replace or waive /build-verification-review. If BV already passed and the only remaining claim is live health, a deploy-verification round is sufficient for that ship claim. Under honesty.enabled: true, a skill pass MUST append or confirm a Mode C–matching entry (ok honesty-status --deploy-health … under require exits 0).
  4. When honesty.enabled: false, ledger append is skipped; D1–D8 still require claims↔record honesty in the review text.
  5. Hard stops unchanged: no merge to {{vcs.git.main_branch}} on findings/blocked; uncited findings invalid; no Tier-3 automation.

Verdicts

VerdictMeaningNext step
passShip/live-health claim honest; Mode C match when requiredAllow DONE wording that asserts shipped/healthy
findingsCited gapsFixer addresses citations only; re-run this skill
blockedEscalating category or fundamental boundary violationHuman required

Escalation categories: security, irreversible, real_money, gates_tier3.

Loop (bounded)

  • Max rounds: 5
  • Reviewer → findings → fixer (cited lines only) → re-review until pass or stop
  • Do not mark ship/DONE under require until pass

Output format

## Deploy verification — <phase-id> round <N>

**Verdict:** pass | findings | blocked
**Frozen spec:** <path>
**Health record:** <path or CI artifact id>
**Mode C:** matched_entry_hash=<…> | missing | honesty disabled

### Findings
| ID | Sev | path:line | Claim vs reality |
| --- | --- | --- | --- |
| DV1 | MAJOR | docs/HANDOVER.md:12 | Claims production healthy; no deploy_health artifact |

### Honest summary (only if pass)
One paragraph: what health record was hashed, what ref was recorded, and that the kit did not probe.

CLI helpers (non-substitute)

./cli/ok honesty-status --deploy-health "<PHASE_ID>" [--frozen-spec PATH] --json
# exit 34 + missing_deploy_health when require_deploy_health: require and no Mode C match

Hard stops

  • Kit never deploys; never opens HTTP(S) to a production URL as verification
  • ref is opaque — never fetched, never executed
  • No merge to {{vcs.git.main_branch}} on findings or blocked
  • Uncited findings invalid — re-review
  • Do not waive /build-verification-review for Auto DONE

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.