agentleFS
Rafaelrr5Rafaelrr5/agent-cortex/skills/delivery-verification/SKILL.md

delivery-verification

Skill1 starsChanged 37 days agoMITView on GitHub

Use when closing or resuming multi-step work. Verify acceptance before declaring delivery.

The file

We split this skill into the rules it gives an agent

6 rules. About 0 are common, 5 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈5

    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
  1. Distinctive:

    Read the artifact and run nonempty checks against the actual acceptance criteria.

    line 17
  2. Distinctive:

    If a baseline, source or criterion is missing, mark that criterion inconclusive.

    line 18
  3. Distinctive:

    Do not turn absent telemetry into zero failures.

    line 19
  4. Distinctive:

    After an external write, read back the exact target.

    line 20
  5. Distinctive:

    Never include secrets in evidence.

    line 30
  6. Very specific:

    Run python scripts/check_delivery.py report.json from this skill directory.

    line 32

≈ “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

Delivery verification

Extract a contract before executing: objective, allowed and forbidden scope, owner, source of authority, observable acceptance criteria, evidence, pending decisions, and next step. Keep one canonical editable handoff, not two competing queues.

Decision rules

  • Local reversible changes require scope authorization; publication, spending, destructive actions and external sends require specific approval.
  • A worker report, completed status or successful process exit is a claim, not proof. Read the artifact and run nonempty checks against the actual acceptance criteria.
  • If a baseline, source or criterion is missing, mark that criterion inconclusive. Do not turn absent telemetry into zero failures.
  • After an external write, read back the exact target. Local tests do not prove deployment.
  • Resume from the handoff and validate local reality; stale notes are not installed state.
  • Permission failures are blockers, not invitations to bypass controls. Retry operational failures only with a reason and a bounded budget.

Evidence contract

Each criterion needs an ID, an observed result, evidence and a verdict: verified, failed, or inconclusive. Completion requires every criterion verified and no pending items. Evidence should identify command, exit code, artifact and revision where relevant. Never include secrets in evidence. Report unknowns explicitly.

Run python scripts/check_delivery.py report.json from this skill directory. The checker validates structure and recorded verdicts only: it cannot establish that someone ran the command or that evidence is truthful. Independently inspect evidence.

Minimal report:

{"criteria":[{"id":"smoke","status":"verified","evidence":"python -m unittest: exit 0; smoke assertion passed"}],"pending":[]}

Run tests: python -B -m unittest discover -s tests -v.

Closure

Return the artifact or usable link, verified results, remaining limitations and the next owner/action. Do not say an attachment was sent unless the outgoing message actually contains it. Save procedure separately from project-specific decisions.

Provenance

Adapted from the supplied local workflow-delivery procedure. No named human author or license was declared in that source; no third-party authorship or license is inferred. This rewrite removes private orchestration and storage conventions.

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.