close
Skill0 starsChanged 28 days agoMITView on GitHub
Close a workflow record properly — rot-lint it, verify the review
26 other projects use a version of this.
Adopted by
Most recently taken up by
- matthewwong525/WongStacktheir own version, first seen Nov 8, 2026
- vzhukovs/podman-desktop-kittheir own version, first seen Nov 1, 2026
- sophonfinance-wq/finance-automation-portfoliotheir own version, first seen Nov 1, 2026
- PapiScholz/session-closuretheir own version, first seen Oct 17, 2026
- GusevV1987/ai-workspace-kittheir own version, first seen Oct 15, 2026
- charly-vibes/incitacionestheir own version, first seen Oct 14, 2026
Used with
Projects with it also use open, reconcile and 6 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.
BlockSkillopen
Open markdown documents in Penpal for collaborative review. Use when you want a human to review a document, plan, or research artifact.
Also in 5 projects that use this
howellsSkillreconcile
Audit a project against current plugin conventions and offer to update structure.
Also in 3 projects that use this
Microsoft AzureSkillvalidate
Run the full CI validation workflow locally before submitting a PR. Use this as a final check to ensure all Azure DevOps CI pipeline steps will pass.
Also in 4 projects that use this
BlockSkillplan
Draft a TDD, PRD, or lightweight plan in the conversation without committing it.
Also in 8 projects that use this
OpenAISkillbootstrap
Project bootstrapping orchestrator for repos that depend on Vercel-linked resources (databases, auth, and managed integrations). Use when setting up or repairing a repository so linking, environment provisioning, env pulls, and first-run db/dev commands happen in the correct safe order.
Also in 3 projects that use this
genomewalkerSkillresume
Resume a thread by loading its ~800-token context capsule
Also in 3 projects that use this
Other versions26 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 26
- Kanevry/session-orchestrator★ 53same name
- wrg32786/aigent-os★ 25same name
- sophonfinance-wq/finance-automation-portfolio★ 16same name
- maxwellsdm1867/wheeler★ 12same name
- matteo-stratega/claude-cortex★ 8same name
- matthewwong525/WongStack★ 7same name
- snapsynapse/skill-provenance★ 7same name
- BriGadja/iapreneurs-claude-code-kit★ 6same name
- Magnus-Gille/claude-skills★ 5same name
- apideck-libraries/api-skills★ 4same name
16 more not listed
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
The file
We split this skill into the rules it gives an agent
2 rules. 0 are common, 2 are distinctive, and 0 are very specific to this project.
0
Common
Most of the 21 edited versions we read keep these.
2
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
- Distinctive: line 119
Never set a terminal status by editing the record.
- Distinctive: line 120
Never close over unreconciled lint findings without the user's explicit say-so (record the override if they insist).
“Common” is measured: kept by at least 6 in 10 of the 21 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.
Files
Close a Workflow Record
Resolve the record: $1 as a slug → perdure/workflows/<date>-<slug>.md;
no argument → the sole in-progress workflow, else ask.
The closer sets the status, in every mode. Steps 1 to 5 prepare the record;
step 6 runs close-workflow.py, which flips the status. Never set completed
or abandoned by editing the frontmatter: a hand-set status skips the
canonical status and updated: stamp, the recorded skip or Outcome line, and
the page a project gets at its first close. Measured 2026-09-17: three of five
headless closes were hand edits, and each left those behind. The close-gate
hook names a hand-close when it sees one and sends you here.
(${CLAUDE_PLUGIN_ROOT} is the plugin directory, set by Claude Code. Without
it, the same scripts sit in the plugin checkout's root scripts/ directory (a
Codex install keeps the whole checkout), in scripts/ beside this SKILL.md (a
skills-CLI install), or in your clone of github.com/arbelh/perdure.)
1. Rot lint first
python3 "${CLAUDE_PLUGIN_ROOT}/scripts/rot-lint.py" <workflow-file>Reconcile anything it prints before proceeding (advisory output — contradicting lines side by side; status-vs-git flags need a human eye). Corrections follow the convention's Correcting a record rules: fix status headers in place with a short dated note; paraphrase retracted claims, never re-quote them; time-scope surviving negatives. A record must not go terminal while its status lies.
2. Verify the review gate
The # Review section must carry a verdict block (schema in
docs/RECORDS-CONVENTION.md, also at references/RECORDS-CONVENTION.md
beside this SKILL.md in a skills-CLI install) or an explicitly recorded skip
with a reason. Do not ask which:
- The task changed files: run the review gate,
/perdure:verdict <slug>, before closing. With no reviewer available, the closer's--skip-review "<reason>"writes the convention's recorded skip and closes on it. - The task changed none (planning only, a dropped plan): record a conscious
skip with its reason,
--skip-review "<reason>". For a dropped plan,--abandon "<why>"alone is enough: the closer records the abandon reason as the skip.
The skip becomes part of the record.
3. Decisions that change an earlier record
For each line under # Decisions: does it change a decision recorded in an
earlier record? If so, name that record's date-slug on the line and say why
(changes 2026-05-12-cache-warmup: TTL 300 → 60, because …). If the earlier
decision is an ADR, supersede it with /perdure:decide instead. Two records
that disagree and name neither are unresolved for every later reader; naming
is what makes the change as loud as the original.
4. Promotion pass
Promote a decision when a task that never reads this record must still obey
it — a default, a name, a boundary, a rule — and it can be stated in one
sentence without this task's filenames. Size does not matter: a one-line
default other tasks inherit qualifies; a ten-file refactor's local choices do
not. Apply the auto-promote-decisions preference: none lists the decisions
and asks; tagged promotes the ones marked {promote} and nothing else;
heuristic (the default) promotes every marked decision and, in addition, any
unmarked one that passes this test, so a mark always wins over the test; all
promotes everything. Next to each promoted decision write whether the mark or
the test promoted it. Two homes: an ADR via /perdure:decide <slug> when the
why and the alternatives matter, marking the decision [promoted → #NNN]; or,
when only the rule matters, propose a one-line addition to the project's
instruction file (CLAUDE.md or AGENTS.md) for the user to place — perdure
never writes outside perdure/.
5. Fill # Outcome
One short paragraph: what shipped or changed, where it landed (commits, tags,
deploys), and a one-line cost estimate when known
(Cost: ~$X, ~N dispatches (estimate)). The closer can append it for you
with --outcome "<text>".
6. Run the closer
python3 "${CLAUDE_PLUGIN_ROOT}/scripts/close-workflow.py" <slug> [--abandon "<reason>"] [--skip-review "<reason>"] [--outcome "<text>"]It refuses on an empty # Review (exit 4) unless --skip-review records the
skip (or, with --abandon, the abandon reason is recorded as it), refuses on
lint findings (exit 3; --force-lint overrides, record why), sets the
canonical terminal status, stamps updated:, appends the abandon
reason or outcome, and creates the HANDOFF page at a project's first close or
regenerates it after. On a record that is already terminal it verifies (lint,
# Review present) and creates or refreshes the page, exit 0; with # Review
empty it refuses on the lint's finding (exit 3), with the section absent on the
gate (exit 4), in both cases unless --skip-review records the skip, the one
write it makes to a terminal record; --outcome is ignored there. Projects that do
not want the page set handoff-regen-at-close: false in
~/.claude/perdure-preferences.json. Never hand-edit the page.
7. Report
One line to the user: record path, final status, verdict or recorded skip, anything withheld or deferred, and the follow-up workflow slug if blockers were deferred.
Safety rules
- Never set a terminal status by editing the record; the closer sets it.
- Never close over unreconciled lint findings without the user's explicit say-so (record the override if they insist).
- Terminal records are immutable except
supersedes:/continued-by:links and a decision renumbering (the decide skill's step 5) — continuing work gets a NEW record. - This skill edits only the workflow record and the derived HANDOFF page.
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.

