agentleFS
thoitiettxl-cyberthoitiettxl-cyber/Claw-II/skills/verification-before-completion/SKILL.md

verification-before-completion

Skill2 starsChanged 7 months agoMITView on GitHub

Standardize the mandatory verification step before claiming a task is done, fixed, pass, deployed, or ready in OpenClaw-Module Android; require grounding on command output, runtime evidence, artifact checks, and explicit remaining-risk notes instead of unsupported assertions. Use when about to report done, fixed, verified, pass, ready to ship, ready to merge, or when a completion gate is needed before concluding.

189 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use chinese-code-review, chinese-commit-conventions 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.
  • Leodorareluctant259Skill

    chinese-code-review

    中文代码审查规范——在保持专业严谨的同时,用符合国内团队文化的方式给出有效反馈

    Also in 8 projects that use this

  • Leodorareluctant259Skill

    chinese-commit-conventions

    中文 Git 提交规范 — 适配国内团队的 commit message 规范和 changelog 自动化

    Also in 8 projects that use this

  • obraSkill

    diagnosing-superpowers

    Use when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness.

    Also in 20 projects that use this

  • jh941213Skill

    docs-ci

    문서 CI(링크·OpenAPI lint·breaking 게이트·mermaid 검증)나 문서 드리프트 감지(docs.yaml 신선도 검사)를 설치하거나 돌릴 때 사용. 문서 내용 작성은 docs-architecture·docs-interfaces·docs-manuals.

    Also in 3 projects that use this

  • jh941213Skill

    docs-interfaces

    API·이벤트 인터페이스 문서(OpenAPI·swagger·AsyncAPI 명세, 엔드포인트 레퍼런스, API 구성도·시퀀스, API CHANGELOG)를 만들거나 코드와 맞출 때 사용. 시스템 구성도는 docs-architecture, API 설계 원칙은 api-design-principles.

    Also in 3 projects that use this

  • +3 more from Leodorareluctant259
  • +4 more from jh941213

The file

We split this skill into the rules it gives an agent

8 rules. 0 are common, 6 are distinctive, and 2 are very specific to this project.

  • 0

    Common

    Most of the 177 edited versions we read keep these.

  • 6

    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 relevant verification action is possible yet.

    line 53
  2. Distinctive:

    If evidence is weak, downgrade the claim instead of upgrading the confidence.

    line 56
  3. Distinctive:

    If key checks cannot be run, state that completion is partial/unverified.

    line 57
  4. Distinctive:

    Prefer direct output over assumptions.

    line 65
  5. Very specific:

    After any state-changing cluster (config/file/service), if full verification is still running or blocked, immediately send a PARTIAL status …

    line 74
  6. Very specific:

    When verification finishes, send a follow-up final verdict (PASS|PARTIAL|FAIL) tied to evidence.

    line 76
  7. Distinctive:

    No executable verification path:

    line 83
  8. Distinctive:

    do not claim completion; report failure class.

    line 88

“Common” is measured: kept by at least 6 in 10 of the 177 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

Verification Before Completion

Indexed Reference Map

  • @[SKILL.md#L1-L95] Frontmatter, trigger rules, SOP, output contract.
  • @[references/INDEX.md#L1-L17] Reading order and reference inventory.
  • @[references/verification-ladder.md#L1-L10] Levels of acceptable verification.
  • @[references/completion-claims.md#L1-L11] How to phrase completion claims safely.
  • @[references/evidence-minimum.md#L1-L10] Minimum evidence required before claiming done.
  • @[references/examples.md#L1-L16] Example completion verdicts.

Scope

  • In-scope:
    • Final verification before saying fixed/done/pass.
    • Turning raw command output into a bounded completion verdict.
    • Recording unverified residual risks explicitly.
  • Out-of-scope:
    • Blindly trusting code changes without execution or evidence.
    • Inflating confidence beyond what checks support.
    • Treating “should work” as equivalent to verified.

KB Mode Notes

  • Reading order:
    1. references/INDEX.md
    2. references/verification-ladder.md
    3. references/completion-claims.md
    4. references/evidence-minimum.md
    5. references/examples.md
  • Coverage Matrix:
    • Verification depth -> verification-ladder.md
    • Safe wording -> completion-claims.md
    • Evidence floor -> evidence-minimum.md
    • Example verdicts -> examples.md
  • Portability note:
    • “Verification” means grounded evidence from actual checks, not confidence language or inferred success.

Trigger Rules

  • Positive:
    • User asks whether something is done, fixed, verified, or safe to conclude.
    • The assistant is about to report completion after edits/config/runtime changes.
    • Prompt contains intent like: double-check before done, verify this, can we claim fixed, are we ready, final check.
  • Negative:
    • Work is still in planning/brainstorming stage.
    • No relevant verification action is possible yet.
    • User explicitly asks for speculative design only.
  • Escalation:
    • If evidence is weak, downgrade the claim instead of upgrading the confidence.
    • If key checks cannot be run, state that completion is partial/unverified.

SOP

  1. List the completion claim
    • What exactly is being claimed: built, fixed, enabled, deployed, passing, etc.
  2. Map claim to checks
    • Choose the commands/artifacts that would actually prove it.
  3. Run or inspect evidence
    • Prefer direct output over assumptions.
    • For plugin shadow claims, evidence must include the resolved install/discovery path, not just profile/config isolation.
  4. Evaluate verification depth
    • Smoke test, targeted verification, or end-to-end proof.
    • --profile by itself is not end-to-end proof of plugin filesystem isolation.
    • Confidence, persona, and polished wording must remain subordinate to evidence; if evidence is incomplete, state uncertainty explicitly instead of compressing to a decisive completion claim.
  5. Record remaining gaps
    • Anything not checked stays explicit.
  6. Mid-flight reporting for state-changing work (mandatory)
    • After any state-changing cluster (config/file/service), if full verification is still running or blocked, immediately send a PARTIAL status instead of staying silent.
    • PARTIAL must include: completed mutations, evidence already collected, exact next verification checks, rollback path, and current risk.
    • When verification finishes, send a follow-up final verdict (PASS|PARTIAL|FAIL) tied to evidence.
  7. Phrase verdict precisely
    • Match wording to evidence strength.
  8. Only then claim completion
    • No “done” before evidence exists.

Error Handling

  • No executable verification path:
    • say the change is prepared, not verified.
  • Output ambiguous or partial:
    • downgrade to partial/pass-with-risk.
  • Check failed:
    • do not claim completion; report failure class.
  • Verification too expensive/risky now:
    • state what was skipped and why.

Output Contract

  • Deliverable:
    • Claim: exact thing being asserted.
    • Checks Run: commands/artifacts inspected.
    • Observed Evidence: real outputs/findings.
    • Remaining Gaps: anything not verified.
    • Completion Verdict: PASS / PARTIAL / FAIL.
  • Validation:
    • A completion claim must be traceable to evidence.
  • Compliance Verdict:
    • PASS when wording matches evidence depth; otherwise PARTIAL or FAIL.

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.