# Agent Skill Bundle — Evidence-first Review

> Generated from the canonical public Skill source. This bundle is an invocation artifact, not a separate editable source.

## Agent instruction

Apply this Skill to the user’s actual task. Follow its inputs, behavior, guardrails, outputs, and verification criteria. Do not merely explain or summarize the Skill. If required task context is missing, ask only for the missing information that blocks correct execution.

## Skill identity

- source: skills/evidence-first-review
- canonical human page: https://wbeen-personal-kb.vercel.app/skills/evidence-first-review/
- version: 0.1.0
- workflow phase: verify
- invocation: either

## Skill contract

Review AI-assisted work against observable evidence instead of accepting a plausible completion claim.

### Inputs

- Claimed result or completion report
- Acceptance criteria or intended outcome
- Available evidence such as source files, tests, screenshots, logs, URLs, or system state

### Outputs

- Verified claims
- Unverified or failed claims
- Evidence references
- Residual risk and recommended next action

### Guardrails

- Never upgrade an unobserved claim to verified.
- Distinguish source evidence from inference.
- Stop and report when required verification access is unavailable.

### Verification

- Every approved claim points to observable evidence or an explicit check result.
- Failed or unavailable checks remain visible in the final review.
- The next action is clear without rereading the full work log.

### Source instructions

## What it does

Evidence-first Review changes the review question from “does this sound finished?” to “what can I actually verify?” It is designed for AI-assisted work where confident completion language can arrive before the underlying state has been checked.

## When to reach for it

Use it at a verification or approval boundary: after implementation, after a migration, before a release, or whenever an agent reports that a task is done.

## What one run does

1. Extract the material claims from the completion report.
2. Map each claim to an observable check.
3. Inspect the available evidence.
4. Mark each claim as verified, failed, or unresolved.
5. Report residual risk and the smallest next action.

## Common failure modes

- Treating a successful command as proof of the entire user-visible outcome.
- Repeating an agent’s own completion summary as evidence.
- Hiding unavailable checks inside a generally positive verdict.
- Expanding scope during review instead of separating follow-up work.

## It’s working if

A reviewer can see exactly why a claim is accepted, what remains uncertain, and what should happen next without reconstructing the whole session.

## Where it fits

This is a **Verify**-phase skill. It pairs naturally with a scoped implementation workflow. A public-safe publishing skill may become a downstream step once that capability is actually published.

## Completion rule

Return the useful task result first. Then state briefly which verification criteria passed, which could not be checked, and any stop/escalation condition that remains.