Evidence Before Claims
Use immediately before claiming work is complete, fixed, passing, committed, published, or ready for handoff; require fresh verification evidence matched to the claim.
Install
npx skills add sandbaseai/workbuddy-skill --skill evidence-before-claimsgit clone --depth 1 https://github.com/sandbaseai/workbuddy-skill.git /tmp/workbuddy-skill
mkdir -p ~/.claude/skills
cp -r /tmp/workbuddy-skill/skills/evidence-before-claims ~/.claude/skills/evidence-before-claimsPersonal skills live in ~/.claude/skills. Use .claude/skills in a project folder to share a skill with your team.
From the SKILL.md
# Evidence Before Claims Core principle: evidence before claims, always. No completion claim without fresh verification evidence. A green-looking diff, a previous run, an Agent report, or confidence is not proof of the current state. ## Apply the gate Before any statement or action that implies success, including committing, opening or merging a pull request, publishing a release, handing off a fix, or moving to the next task: 1. **Identify** the exact claim and the command, query, or artifact that proves it. 2. **Run** the full check now, at the scope of the claim. Do not substitute a cheaper partial check. 3. **Read** the complete relevant output, exit code, failure count, warnings, and environment assumptions. 4. **Verify** whether the evidence actually supports the claim. If not, report the actual status and gap. 5. **Only then** make the narrowest claim supported by the evidence or perform the authorized next action. Do not use “should”, “probably”, “looks good”, or an equivalent confidence signal as a substitute for verification. If a check cannot run, say `not verified`, explain why, and preserve the smallest safe follow-up. ## Match claims to proof | Claim | Fresh evidence required | Insufficient evidence | |---|---|---| | Tests pass | complete relevant test command and zero failures | lint only, old output, one focused test | | Build succeeds | actual build command exits 0 | syntax check or packaging intent |
Details
- Skill name
evidence-before-claims- Repository
- sandbaseai/workbuddy-skill
- Path
skills/evidence-before-claims/SKILL.md- License
- MIT
- Repo updated
- 8 September 2026
FAQ
What does the Evidence Before Claims skill do?
Use immediately before claiming work is complete, fixed, passing, committed, published, or ready for handoff; require fresh verification evidence matched to the claim.
How do I install the Evidence Before Claims skill?
skills CLI (Claude Code, Cursor, Codex and more): npx skills add sandbaseai/workbuddy-skill --skill evidence-before-claims. Skills go in ~/.claude/skills (personal) or .claude/skills (project), and Claude loads them automatically when a task matches the description.
Is Evidence Before Claims popular?
It lives in sandbaseai/workbuddy-skill, which has 9 GitHub stars and 3 forks.
Related skills
Access Control Security Best Practices
9sandbaseai/workbuddy-skill
Design, review, implement, and verify least-privilege access controls across users, services, tenants, resources, and administrative workflows.
Accessibility Review
9sandbaseai/workbuddy-skill
Use when reviewing a web interface or interaction for semantic, keyboard, focus, screen-reader, ARIA, and contrast accessibility risks.
Acquire Codebase Knowledge
9sandbaseai/workbuddy-skill
Use when the user asks to map, document, or onboard into an existing codebase and needs evidence-backed architecture, structure, conventions, integrations, testing, and concerns documentation.
Incident Postmortem
9sandbaseai/workbuddy-skill
Use after a production outage, significant degradation, data incident, or near miss to reconstruct evidence, quantify impact, identify systemic causes without blame, and create owned, dated prevention actions.
Evaluation Methodology
9sandbaseai/workbuddy-skill
Use when measuring WorkBuddy Skill quality, interpreting evaluation results, calibrating thresholds, or deciding which triggering, output, robustness, or orchestration gap to improve first.
Incident Runbook Templates
9sandbaseai/workbuddy-skill
Use when creating or reviewing an incident runbook with detection, triage, authorized mitigation, escalation, verification, rollback, and communication boundaries.
Browse the most popular Claude skills, all other skills, or write your own with the skills generator.