verify-ticket-closure
Skill4 starsChanged 4 months agoApache-2.0View on GitHub
Independently certify whether one exact immutable candidate satisfies its frozen deterministic ticket contract; never redefine the ticket, Spec, or CCL universe.
In this repository51 other files it gives its agents.
- AGENTS.mdAGENTS.md
- architecture-remediationSkill
- attentionSkill
- classify-docSkill
- codebase-designSkill
- codebase-memory-mcpSkill
- codegraphSkill
- coding-standardsSkill
- contract-ledgerSkill
- conventional-commitsSkill
- database-migrationsSkill
- deduplicate-codeSkill
- diagnosing-bugsSkill
- domain-modelingSkill
- format-codeSkill
- github-issue-dependenciesSkill
- graphifySkill
- grillingSkill
- grill-meSkill
- grill-with-docsSkill
- handoffSkill
- implement-ticketSkill
- improve-codebase-architectureSkill
- project-delivery-managementSkill
- project-trackingSkill
- prototypeSkill
- repowiseSkill
- researchSkill
- resolving-merge-conflictsSkill
- review-architectureSkill
- review-specSkill
- setup-matt-pocock-skillsSkill
- spec-contractSkill
- spec-merge-cleanupSkill
- tddSkill
- teachSkill
- to-adr-docSkill
- to-docSkill
- to-specsSkill
- to-ticketsSkill
- triageSkill
- verify-architectureSkill
- verify-codeSkill
- verify-spec-closureSkill
- verify-specSkill
- verify-ticket-decompositionSkill
- wayfinderSkill
- wiki-lintSkill
- wiki-syncSkill
- wiki-synthesizeSkill
- writing-great-skillsSkill
The file
We split this skill into the rules it gives an agent
9 rules. About 0 are common, 8 are distinctive, and 1 is 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.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 12
Before executing this skill, read both of these files in full:
- Distinctive: line 24
All retained frozen-contract, proof-independence, per-cell proof, exclusion-guardrail, retry, gap-classification, PASS/FAIL completeness, …
- Very specific: line 41
Before semantic proof, independently fetch current origin/main and require the protected- candidate gate defined by …
- Distinctive: line 100
every finding/falsifier is derived from an assigned cell, applicable exclusion, or valid gap classification rather than an out-of-contract …
- Distinctive: line 106
Do not inspect new candidate evidence, mutate anything, broaden the semantic universe, or invent a new finding during this handshake.
- Distinctive: line 109
If the exact result remains internally consistent, return exactly:
- Distinctive: line 121
If this verifier discovers that its own exact PASS/FAIL is internally contradictory or relied on an out-of-contract semantic assumption, it …
- Distinctive: line 145
Every retained PASS/FAIL result contract is strengthened to include:
- Distinctive: line 163
All other result fields and proof obligations in the retained procedure remain unchanged.
≈ “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.
Files
Verify Ticket Closure
Authoritative procedure composition
Before executing this skill, read both of these files in full:
.agents/skills/verify-ticket-closure/PROCEDURE.md.agents/skills/verify-ticket-closure/POLICY_SYNC_ADMISSION.md
PROCEDURE.md preserves the complete established independent-verifier procedure.
POLICY_SYNC_ADMISSION.md owns the candidate-admission, Execution-Anchor, policy-sync,
Attention-identity, terminal candidate recheck, and result-finalization candidate-identity
mechanics. On those subjects it overrides retained procedure text. This SKILL.md is the
authoritative entrypoint; on every other conflict, this file wins. Failure to load either
retained file is invalid verification.
All retained frozen-contract, proof-independence, per-cell proof, exclusion-guardrail, retry, gap-classification, PASS/FAIL completeness, and non-mutation rules remain normative.
Execution Anchor invocation integrity
A current v3 verification dispatch must receive and bind both:
- Ticket baseline — immutable historical/per-ticket lineage anchor;
- Execution Anchor — exact fetched shared-Spec-branch commit admitted by
$implement-ticketbefore this attempt's substantive mutation.
TICKET_CLOSURE_STATE is the exact policy-sync-aware candidate state relative to the
Execution Anchor. Upstream/prerequisite/shared-branch commits between Ticket baseline and
Execution Anchor are not part of this ticket candidate, and later committed exact-main
workflow-policy synchronization is delivery-tip history rather than candidate mutation only
when the current machine guard admits it.
Before semantic proof, independently fetch current origin/main and require the protected-
candidate gate defined by POLICY_SYNC_ADMISSION.md:
python .agents/skills/implement-ticket/scripts/candidate_guard.py check \
--anchor <execution-anchor> \
--remote origin \
--default-branch mainRequire exact:
WORKFLOW CANDIDATE GUARD: VALID
Protected control-plane paths: 0The verifier must execute this gate itself. It must not rely on a parent-reported guard result. A protected path, inadmissible committed delivery-tip change, stale/non-main policy blob, mixed policy/product commit, dirty protected path, or non-descendant HEAD is invalid verification. The verifier must not repair, restore, suppress, format, or otherwise mutate the candidate; return invalid verification and emit neither PASS nor FAIL.
Mechanically admit the Attention lifecycle state with:
python .agents/skills/implement-ticket/scripts/attention_gate.py verify \
--ticket <ticket-number> \
--contract <ticket-or-remediation-contract.json> \
--baseline <immutable-ticket-baseline> \
--branch <ticket-branch> \
--execution-anchor <execution-anchor> \
--expected-receipt-hash <attention-receipt-hash>Require exact IMPLEMENT TICKET ATTENTION: VALID, exact printed Ticket baseline, exact
printed Execution Anchor, and exact printed candidate state equal to the dispatched
TICKET_CLOSURE_STATE.
Missing/mismatched Execution Anchor, a v1 Attention receipt, candidate state derived from Ticket baseline, candidate mutation, or committed delivery-tip movement not admitted by the policy-aware candidate guard is invalid verification: emit neither PASS nor FAIL. Never mint, repair, or reinterpret an Execution Anchor in the verifier.
Authorized committed exact-main workflow-policy synchronization after the Execution Anchor
is not, by itself, invalid verification. It remains admissible only through the exact machine
predicates in POLICY_SYNC_ADMISSION.md and must be independently rechecked by this verifier.
Result-finalization and verifier self-retraction
A retained PASS/FAIL result is not admissible by $implement-ticket until one exact
result-finalization handshake completes in this same verifier context. The parent supplies
the SHA-256 of the exact PASS/FAIL bytes it received and asks only for result finalization;
it may not argue findings, introduce new evidence, suggest a desired outcome, or ask the
verifier to reinterpret the frozen contract.
During finalization, re-check only the internal consistency of the exact returned result against the already-bound deterministic contract and already-used evidence. Confirm that:
- every finding/falsifier is derived from an assigned cell, applicable exclusion, or valid gap classification rather than an out-of-contract semantic assumption;
- the result does not contain mutually incompatible claims about the same frozen predicate;
- PASS/FAIL completeness and failure-saturation statements are internally consistent with the result's own per-cell and guardrail dispositions.
Do not inspect new candidate evidence, mutate anything, broaden the semantic universe, or invent a new finding during this handshake.
If the exact result remains internally consistent, return exactly:
TICKET CLOSURE RESULT: CONFIRMED
Result SHA-256: <sha256 of exact PASS/FAIL bytes supplied by parent>
Ticket: #<n>
Ticket Contract Hash: <hash>
Attention Receipt Hash: <hash>
Execution Anchor: <sha>
Candidate state: <hash>If this verifier discovers that its own exact PASS/FAIL is internally contradictory or relied on an out-of-contract semantic assumption, it must retract that result itself and return exactly:
TICKET CLOSURE: INVALID
Invalid reason: self-contradictory-verdict
Retracted result SHA-256: <sha256 of exact PASS/FAIL bytes supplied by parent>
Retracted verdict: <PASS|FAIL>
Ticket: #<n>
Ticket Contract Hash: <hash>
Attention Receipt Hash: <hash>
Execution Anchor: <sha>
Candidate state: <hash>
Contradiction: <concise verifier-owned description of the incompatible or out-of-contract claim>INVALID is not PASS or FAIL and carries no implementation finding. After emitting it,
this verifier context is finished and must not emit a replacement semantic verdict. Only
the same verifier context may retract its own result; neither $implement-ticket nor a
human may manufacture this payload on the verifier's behalf.
Frozen result identity
Every retained PASS/FAIL result contract is strengthened to include:
Ticket baseline: <immutable sha>
Execution Anchor: <sha>
Candidate state: <policy-sync-aware hash relative to Execution Anchor>Retry state must preserve the same Execution Anchor unless $implement-ticket legally
abandons the stale attempt before mutation and starts a new pre-mutation attempt from a new
clean exact-remote anchor. A verifier never performs that transition.
Immediately before either terminal PASS or FAIL, fetch current origin/main, rerun the exact
policy-aware candidate guard from POLICY_SYNC_ADMISSION.md against the same Execution
Anchor, and rerun attention_gate.py verify against the same contract, Attention Receipt
Hash, and Execution Anchor. Any identity drift or inadmissible candidate/control-plane state
invalidates the verifier run and emits neither PASS nor FAIL.
All other result fields and proof obligations in the retained procedure remain unchanged.
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
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.

