pm-verify
Skill2 starsChanged 8 months agoMITView on GitHub
Verify acceptance criteria for a phase or entire work package. Runs verification based on each criterion's method (AutomatedTest, Manual, AgentReview) and records results via the MCP tool.
In this repository15 other files it gives its agents.
- pm-auditSkill
- pm-brainstormSkill
- pm-cleanupSkill
- pm-codesmellsSkill
- pm-doneSkill
- pm-exploreSkill
- pm-housekeepingSkill
- pm-implementSkill
- pm-nextSkill
- pm-planSkill
- pm-refine-frSkill
- pm-scaffoldSkill
- pm-statusSkill
- pm-triageSkill
- pm-troubleshootSkill
The file
Show its rules and text
We split this skill into the rules it gives an agent
14 rules. About 0 are common, 9 are distinctive, and 5 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈9
Distinctive · estimate
This author's own choices. Worth a look.
5
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 12
Verify that acceptance criteria for a phase or work package are met.
- Very specific: line 20
--dry-run: If present, show what would be verified without recording results - Very specific: line 22
If no arguments provided, use the AskUserQuestion tool:
- Very specific: line 29
Call mcp__pinkrooster__get_work_package_details with the WP ID (extract from phase ID if needed).
- Distinctive: line 32
All phases with their acceptance criteria
- Distinctive: line 33
If a phase ID was given, focus on that phase only
- Distinctive: line 34
If a WP ID was given, verify all phases with unverified criteria
- Distinctive: line 43
Check if the related tests exist (use Grep/Glob to find test files)
- Distinctive: line 44
Run the relevant test suite:
- Distinctive: line 51
Read the target files associated with tasks in the same phase
- Distinctive: line 53
Check that the described behavior or structure exists
- Very specific: line 67
If --dry-run, skip this step and show what would be recorded.
- Distinctive: line 104
If a criterion name doesn't match any criterion on the phase, report an error
- Distinctive: line 105
Always show the full report even if all criteria pass
≈ “Common” is an estimate: at least 1 in 5 of the 372,429 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 Acceptance Criteria
Verify that acceptance criteria for a phase or work package are met. Records verification results
and timestamps. Does NOT change entity states — use /pm-done to complete phases after verification.
Step 0: Parse Arguments
Parse !arguments for:
- Target ID: phase ID (e.g.,
proj-1-wp-2-phase-1) or WP ID (e.g.,proj-1-wp-2) --dry-run: If present, show what would be verified without recording results
If no arguments provided, use the AskUserQuestion tool:
- Question: "What would you like to verify? Provide a phase or work package ID."
- Header: "Verify"
- Options:
[{label: "Phase", description: "Verify acceptance criteria for a specific phase (e.g., proj-1-wp-2-phase-1)"}, {label: "Work Package", description: "Verify all phases in a work package (e.g., proj-1-wp-2)"}]
Step 1: Load Context
Call mcp__pinkrooster__get_work_package_details with the WP ID (extract from phase ID if needed).
Extract:
- All phases with their acceptance criteria
- If a phase ID was given, focus on that phase only
- If a WP ID was given, verify all phases with unverified criteria
Build a verification queue: list all acceptance criteria that have verifiedAt == null or need re-verification.
Step 2: Verify Each Criterion
For each criterion in the queue, run verification based on verificationMethod:
AutomatedTest
- Check if the related tests exist (use Grep/Glob to find test files)
- Run the relevant test suite:
.NET tests:dotnet testwith appropriate filterDashboard tests:cd src/dashboard && npm test
- Parse test output for pass/fail
- Record result: "PASS: All N tests passed" or "FAIL: N/M tests failed — {details}"
Manual
- Read the target files associated with tasks in the same phase
- Review the code against the criterion's description
- Check that the described behavior or structure exists
- Record result: "PASS: {evidence}" or "FAIL: {what's missing}"
AgentReview
- Analyze the code related to the criterion
- Assess against the criterion's description using code understanding
- Record result: "PASS: {assessment}" or "FAIL: {assessment}"
Step 3: Record Results (unless --dry-run)
For each phase with verified criteria, call mcp__pinkrooster__verify_acceptance_criteria with:
phaseId: the phase IDcriteria: array of[{"Name": "...", "VerificationResult": "..."}]
If --dry-run, skip this step and show what would be recorded.
Step 4: Report
## Verification Report — {targetId}
| Phase | Criterion | Method | Result |
|-------|-----------|--------|--------|
| {phaseName} | {criterionName} | {method} | PASS/FAIL |
| ... |
### Summary
- **Total criteria**: {count}
- **Passed**: {passCount}
- **Failed**: {failCount}
- **Skipped**: {skipCount} (if any)
### Failed Criteria
- **{criterionName}** ({phaseName}): {failureReason}
- ...
### Next Steps
- All passed? → Complete the work: `/pm-done {phaseId or wpId}`
- Failures? → Fix failing criteria: `/pm-implement {relevant-task-or-phase-id}`, then re-verify: `/pm-verify {targetId}`
- Check project progress: `/pm-status`If --dry-run:
"Dry run complete. {count} criteria would be verified. Run without --dry-run to record results."
Constraints
- This skill does NOT change entity states — it only records verification data
- Re-verification is safe (idempotent) — new results overwrite old ones
- For AutomatedTest criteria, always run the actual tests — don't just check if test files exist
- For Manual and AgentReview criteria, provide specific evidence from the code
- If a criterion name doesn't match any criterion on the phase, report an error
- Always show the full report even if all criteria pass
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.

