本文へ移動
cccskills
無料GitHub で公開

os-check-work

ALWAYS invoke this skill when the user asks about work done outside this session - "check work", "check the others", "check other sessions" - or to accept one: "the session is done, check it", "can we merge it" - in any language. You are the receiving party: treat the report as a claim, verify each part against machine state, name every gap between claimed and true. Verified-ready work merges in the same pass. End with what is left, one next step, and any reply another session needs, ready to paste.

インストール方法を見る

含まれるファイル(2)

  • SKILL.md7.4 KB
  • references/01-accepting-a-build-session.md2.3 KB

SKILL.md(原文)

インストールする前に、エージェントに与えられる指示の中身を確認できます。

os-check-work

You are the receiving party: another session says it finished, and your job is to find out whether that is true. Reports run optimistic - they are written by whoever did the work. Triggers live in the description above; one more is taking over work you did not do.

Language

Write in the language the user speaks in this session, detected from the conversation. Code, file names and identifiers stay English.

The stance

  • A report is a claim by the party being assessed, and its adjectives are not measurements: "tests are green" vs the count that ran. Ask for the number.
  • "Done" is a snapshot. Reviews get revoked, branches fall behind and checks flip, often with no new commit. Re-read immediately before you act.

Two modes

Plural is a sweep: every sibling session, one line each, statuses only. Singular - "accept this", a named session or pull request - is an accept: full verification, then acceptance. A sweep that finds verified-ready work rolls into accepting it - the accept steps run first, never skipped. Where your tool cannot list other sessions, sweep the open pull requests instead (gh pr list) and say so.

Sweep

One line per session, one call per pull request - the reader skims these and reads the closing block. No table:

<name a person recognises (#PR)> - <product change>; <all done? plus proof in 2-4 words, or what is missing>; <archive: yes / not yet>.

  • Name from the ticket or PR title, never your shorthand; the change for the product, not a file list.
  • A yes carries its proof ("checks green, review approved"); never a bare yes.
  • "All done" ≠ "can archive": finished work can still await acceptance.
  • Order: needs-the-user, then broken, then still-working - a running session is one word, not a finding.
  • No deep verification in a sweep: say what you did not open and what that leaves unverified. Never announce your working mode - "cheap checks" is bookkeeping.
  • A session waiting on an answer gets it in the same pass - see "The reply"; when that needs the deep read, the reply becomes the do-this-next.
  • Name the sessions you could not read, and why.

Accept - steps 1 to 5 (never per-session inside a sweep)

  1. Find the claim. Read the transcript tail where tools allow - the last message is the report; read it once, do not quote it back. No tools → the pull request is the artefact; say so.

  2. Verify each claim at its source.

    The session claimsWhere the truth lives
    checks are greengh pr checks - counts and states, not the summary word
    it was approved--json reviewDecision,latestReviews - revocable on the same commit
    no open discussionsthe unresolved-thread count, not "I addressed the comments"
    ready to mergemergeStateStatus - and again right before merging
    tests were addedfind them in gh pr diff - the most common false report
    it stayed in scopegh pr diff --name-only vs the task; flag unasked files
    linked to the taskthe ticket in branch, title or body

    Prefer numbers to words. Check what the report did not mention: scope creep and missing tests are never claimed - they are found.

  3. Name the gaps - the deliverable. One line each: Claimed: … Measured: …. No gaps → say so in one line; a clean intake is a real result.

  4. Accept, hand back, or unblock.

    • Content does not hold → hand it back with the reply written (below); never fix it silently - that makes you the author of work you were checking.
    • Mechanical unblocking is yours without asking: update a branch from the main line, restart a stuck check - reversible plumbing, no content change.
    • Merging needs no word at all: work whose claims verified merges in any mode, no round-trip. Exactly two things stop it - a claim that failed, and an instruction on this task that merges happen on command only (an orchestrator may own the merge; that instruction beats any standing policy). The verification itself is never skipped.
  5. Clean up. Remove its leavings - working copy, branch, temp files; ask before deleting what is not certainly its. Archive verdict → closing block.

The reply is part of the pass

When a checked session needs to hear something back - a defect, a question, a decision - write the reply in the same pass; do not leave the translation to the user.

**To <session name>** - paste this:
<the instruction, verbatim>
  • Self-contained - the target never read this chat: name the PR, file, gap.
  • It speaks as the user: a plain instruction, not a review essay.
  • One best answer → one block; a real fork → two, recommended first and marked; past three it is a decision for os-ask-simple.
  • Carry the claimed/measured pair - evidence, not adjectives.

Ending the pass

Both modes end the same way, always last, nothing after it:

| | |
|---|---|
| **Still open** | <what remains, or "nothing"> |
| **Do this next** | <the one action> - <one clause of reasoning> |
| **On your word** | <what you will do once they say it> |
  • Say whether anything is left, including what waits on the user; a remainder past three items goes to os-whats-next.
  • One action, never a choice: what unblocks the most goes first, and the reasoning clause says so.
  • Nothing ready is a real answer - "nothing to accept yet, watch X". Never invent an action; a step that is not yours → say whose (os-step-by-step).
  • Word already given (in the ask, or by standing policy) → the third row reports what you did. Pass closed everything → run os-whats-next.
  • Never end on a question mark - you have the measurements, so the opinion.

Hard rules

  1. Never repair another session's work silently - name it, return it.
  2. Never accept on the strength of the report alone; detail is not evidence.
  3. Nothing irreversible without the word - given in the ask itself or by a standing policy; a task-scoped restriction beats both. Merging verified work already has that policy: step 4.
  4. Re-read state immediately before acting.
  5. Report gaps as measurements: what was claimed, what you found, in that order.
  6. Say what you could not check - no access is a gap, not absence of problems.
  7. End with the closing block; findings without a recommendation are unfinished.

Known gotchas

  • Behind the main line ≠ conflicted: it needs an update, not the author.
  • A conflicted branch may have run no checks: nothing-ran looks green.
  • Delete a branch only from outside its own working copy.
  • Shared state: stashes, containers, databases - do not clean up what another live session is using.
  • A draft is not a finished session; check the author considers it done.

レビュー

まだレビューはありません。使ってみた感想をお寄せください。

同じリポジトリのスキル

概要と使いどころ

ALWAYS invoke this skill before asking the user any technical question or offering options, and whenever they ask to be asked in plain words - "ask simple", "ask me simply", "ask me in plain words" - in any language. ALWAYS invoke it too when the user asks for your view on a technical choice: whether it is worth doing, more than the problem needs, or replaceable by something simpler, and which option you would take. Invoke it even when the code makes the answer look obvious: the checks are what make the answer more than a guess. Rewrites the question in plain words and always ends with one marked recommendation. A structural choice first passes six checks, shown as a table: effort now, simpler substitute, extra work later, lock-in, over-engineering, easy to undo. Doing nothing is always weighed.

日本語の概要は準備中です。原文の説明を表示しています。

kharmanskyi/open-steps1,2812026年10月6日 更新

ALWAYS invoke this skill when the user asks where the project as a whole stands - "where are we", "what's the big picture", "what have we built", "what is in this project", "map the project", "what does this thing even do", "what is stale", "what can we delete", "update the roadmap" - in any language, and whenever a session report has just been written. It keeps one file, BIG-PICTURE.md: what the product is, every feature with how far it got, which parts nobody has touched, and what is queued next. Age and wiring are measured from git; how far a feature got comes from the reports or says "not checked". Where a tracker is connected it offers to open the queue as tickets, after a yes and never before. Never invents work, never deletes code.

日本語の概要は準備中です。原文の説明を表示しています。

kharmanskyi/open-steps1,2812026年10月6日 更新

ALWAYS invoke this skill when work wraps up or the user asks how it went - "done or not", "are we done", "what happened", "report", "what's the status of this ticket" - in any language, and when a Stop hook asks for a session report. Produces a ten-line plain- language report: a lead, a checkmark table, and a verdict - fully done, anything needed from you, new debt, safe to close. Every "yes" names its proof; unverified says "not checked". Saves the report so the next session starts from it instead of re-exploring the repo.

日本語の概要は準備中です。原文の説明を表示しています。

kharmanskyi/open-steps1,2812026年10月6日 更新

ALWAYS invoke this skill when the user asks for simpler or shorter about something said or written - "say it simply", "what does this mean", "I don't understand your answer", "too long", "wait, what?", "bro" - in any language, about any text: your own answer, a report, a review comment, an error. "I don't understand what to DO" is os-step-by-step; this skill restates text. It restates for a reader who does not read code: leads with the point, keeps every number, warning and caveat - no facts added, no bad news dropped. A number returns exactly that many points, most important first.

日本語の概要は準備中です。原文の説明を表示しています。

kharmanskyi/open-steps1,2812026年10月6日 更新

ALWAYS invoke this skill when you need the user to act - run a command, paste a secret, click, approve - and whenever they ask how to do something or say they do not know what to do: "step by step", "walk me through it", "what do I do", "what should I do", "I don't understand what to do", "explain what I need to do", in any language. Picking which task comes next is os-whats-next; this skill is for doing the thing in front of you. First earn the ask: try it yourself, find another route, shrink it to the part only they can do. Then one action per step, commands labelled by what they touch, no jargon. Commands are single lines that prompt for any value - typing hidden for secrets - and confirm in plain words. Afterwards verify their step.

日本語の概要は準備中です。原文の説明を表示しています。

kharmanskyi/open-steps1,2812026年10月6日 更新

ALWAYS invoke this skill before anything hard to undo gets agreed to - a contract, a purchase, a migration, a launch, a price change, a reorganisation - and whenever the user asks "what could go wrong", "what are we missing", "poke holes in this", or for a premortem or a red team, in any language. Assumes the decision already failed and works backwards to find out why, in a fresh agent that had no hand in making it. Sweeps nine areas and shows what each produced, including the empty ones. Ends on one verdict: go ahead, go but fix these first, try it small first, think again, do not do this.

日本語の概要は準備中です。原文の説明を表示しています。

kharmanskyi/open-steps1,2812026年10月6日 更新

kharmanskyi のスキルをすべて見る

このスキルの問題を報告する