Ve evidence auditor
Skill onyx679/automotive-ve-ai-skills-kit/skills/ve-evidence-auditor
AI Skill templates and workflow scoring tools for automotive value engineering and VAVE productivity scenarios
npx -y skills add onyx679/automotive-ve-ai-skills-kit --skill ve-evidence-auditorAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
What its author says it does
Copied from the file, not written here
Audit value-engineering AI Skill claims against concrete evidence links, review status, and honest wording boundaries. Use when preparing resumes, portfolios, project reports, best-practice documents, release notes, or management summaries that must distinguish verified work, submitted PRs, pending reviews, assumptions, and claims that should not be made.
SKILL.md
3.1 KB, as published. Nobody here has run it
VE Evidence Auditor
Use this skill to turn project claims into an evidence-backed claim matrix before publishing them in a resume, portfolio, report, or internal best-practice document.
Inputs
Accept any of:
- Resume bullets or project descriptions.
- GitHub repositories, releases, pull requests, issues, CI runs, or portfolio pages.
- Work logs, adoption reports, Skill outputs, SOPs, and demo artifacts.
- A CSV claim ledger with
claim,source_kind,evidence_level,verdict,evidence_status,evidence_url,resume_wording, andboundary.
Workflow
- Break the text into atomic claims. Each claim should assert one thing only.
- Attach a concrete evidence URL or file path to every claim.
- Grade the source:
source_kind:platform-record,repository-artifact,third-party,self-reported, ormixed.evidence_level:L4for directly checkable platform/repository records;L3for multiple independent sources;L2for one credible third-party source;L1for self-reported or weak evidence.verdict:confirmed,largely-credible,doubtful, ordebunked.
- Mark evidence status:
verified: public link or local artifact proves the claim.open: submitted but not merged, approved, or accepted.pending-review: waiting for maintainer, manager, or business review.missing: no evidence yet.
- Classify claim level:
resume-ready: verified evidence exists and wording is not overstated.boundary-only: evidence exists, but wording must mention open or pending status.do-not-claim: evidence is missing, wording implies unearned completion, or the claim depends on confidential/internal facts not available.
- Rewrite risky claims into safe wording.
- Keep a "do not claim yet" section for tempting but unsupported statements.
Output
Return:
# Evidence Claim Matrix
## Summary
- Resume-ready:
- Boundary-only:
- Do-not-claim:
## Claim Review
| Claim | Source | Evidence level | Verdict | Status | Claim level | Evidence | Safe wording | Boundary |
|---|---|---|---|---|---|---|---|---|
## Do Not Claim Yet
- <claim>: <reason>
Quality Checks
- Never upgrade
openorpending-reviewPRs into merged contribution claims. - Never claim internal deployment, supplier data handling, or business impact unless there is approved evidence.
- Prefer "submitted PR", "published demo", "verified by CI", and "review-pending" over vague prestige wording.
- Treat self-reported or marketing-only evidence as
L1unless it is backed by a directly checkable platform record. - If a claim has no URL, artifact, or owner-verifiable source, keep it out of the resume.
- Preserve the difference between public portfolio evidence and real enterprise work.