maintain-verification-skill
Check and repair a project verification skill against source and live behavior.
Skill by wcygan
Install
Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.
npx agentmarket install skill/wcygan-maintain-verification-skill
Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.
Get the weekly diff
The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.
About
Reviews
Write a reviewNo reviews yet. Installed it? Reviews come from verified installers only. How reviews work
Permissions
What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.
- Tools
- No tools found
- Secrets
- None asked for
- Install scripts
- None found
Network hosts it talks to 1
codeload.github.com· found in install command
Files it touches
~/.claude/skills/maintain-verification-skill
What's inside
- Skill name
maintain-verification-skill- Description
- Check and repair a project verification skill against source and live behavior. Use when its commands, journey map, or evidence requirements may have drifted.
- License
- MIT
- Metadata
- kind: component · maintenance: local · upstream-path: pstack/skills/maintain-verification-skill · upstream-author: Lauren Tan · upstream-revision: 1e5c90a4b6f249917bd21fe361b3a2a7deb741b8 · upstream-repository: https://github.com/cursor/plugins
- Instructions
- 26 lines in SKILL.md · view on GitHub
Files (2)
- SKILL.md
- LICENSE.txt · references/
SKILL.md (first lines)
# Maintain a verification skill 1. Locate the intended verification skill and feature map. Resolve multiple targets from the request; if none exists, report that creation is needed. 2. Reconcile the index with its entries and current source. Check every mapped feature, and look for missing user surfaces with concrete source evidence. 3. Exercise every mapped feature on the real surface, even if source looks unchanged. Use one driver per shared instance or isolated sessions. Health-check before use and after surprising failures; reset unusable state instead of continuing blindly. 4. Classify failures: documentation drift, a harness gap, a product regression, or an unavailable prerequisite. Fix only the verification skill and its owned helpers. Report product regressions without rewriting expected behavior to hide them. 5. Rerun corrected instructions live. For a health-check failure caused by skill drift, correct it and retry once before reporting the remaining blocker. Preserve evidence and clean owned resources after each failed attempt and final teardown. 6. Report clean, changed, or blocked with per-feature coverage and evidence. An unreachable feature needs its attempted route and missing prerequisite recorded; it is not a passing feature. Publish corrections only if requested. Done when every feature has a source and runtime disposition, corrections have been exercised, and surviving evidence supports the outcome. Incomplete live coverage keeps the overall outcome blocked even when some corrections succeeded. Adapted from pstack; [license and attribution](references/LICENSE.txt).
From wcygan/agent-skills, MIT.
More from wcygan/agent-skills
- 0.570score
coordinate-work
wcygan
Assign independent units, control shared writes, and reconcile results.
SkillAFreshness: AClaude Code +9 - 0.570score
create-verification-skill
wcygan
Create a project skill that launches, exercises, and verifies real application behavior.
SkillAFreshness: AClaude Code +9 - 0.570score
compare-solutions
wcygan
Compare competing solutions and verify a coherent synthesis.
SkillAFreshness: AClaude Code +9 - 0.569score
implement-change
wcygan
Implement a bounded feature or refactor in verifiable steps.
SkillAFreshness: AClaude Code +9 - 0.567score
design-options
wcygan
Compare designs and sketch interfaces before implementation.
SkillAFreshness: AClaude Code +9 - 0.567score
explain-code
wcygan
Explain subsystem behavior and ownership from evidence.
SkillAFreshness: AClaude Code +9 - 0.567score
explain-history
wcygan
Explain the evidence behind a design decision or past failure.
SkillAFreshness: AClaude Code +9 - 0.567score
improve-workflow
wcygan
Find evidence-backed improvements to an engineering process.
SkillAFreshness: AClaude Code +9 - 0.567score
decision-log
wcygan
Keep an auditable record of consequential choices and evidence.
SkillAFreshness: AClaude Code +9 - 0.567score
deliver-change
wcygan
Present or publish a verified result within authorized scope.
SkillAFreshness: AClaude Code +9 - 0.567score
design-loop
William Cygan
Design, style, or redesign websites and apps through investigation, layout, design-system creation, implementation, validation, and delivery.
SkillAFreshness: AClaude Code +9 - 0.565score
discover-ambitious-goals
William Cygan
Find valuable project outcomes and stretch goals from product and engineering evidence.
SkillAFreshness: AClaude Code +9
Signals
- Stars
- 0
- Forks
- 0
- Open issues
- 0
- Last commit (this folder)
- Oct 14, 2026 (14 days ago)
From GitHub (fetched Oct 24, 2026).
Details
- License
- MIT
- Repository
- github.com/wcygan/agent-skills
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
More from wcygan
- 0.570score
create-verification-skill
wcygan
Create a project skill that launches, exercises, and verifies real application behavior.
SkillAFreshness: AClaude Code +9 - 0.570score
compare-solutions
wcygan
Compare competing solutions and verify a coherent synthesis.
SkillAFreshness: AClaude Code +9 - 0.570score
coordinate-work
wcygan
Assign independent units, control shared writes, and reconcile results.
SkillAFreshness: AClaude Code +9 - 0.569score
implement-change
wcygan
Implement a bounded feature or refactor in verifiable steps.
SkillAFreshness: AClaude Code +9 - 0.567score
decision-log
wcygan
Keep an auditable record of consequential choices and evidence.
SkillAFreshness: AClaude Code +9 - 0.567score
deliver-change
wcygan
Present or publish a verified result within authorized scope.
SkillAFreshness: AClaude Code +9
Cite this page
maintain-verification-skill. Agent Mart. https://agentmart.dev/skills/wcygan-maintain-verification-skill (last updated 2026-10-24)