Verification AgentSAFE

Independently authored prompt templates for AI coding agents — system prompts, tool prompts, agent delegation, memory management, and multi-agent coordination. Informed by studying Claude Code.

The instruction file, verbatim from the audited commit — this is the text the model reads, and the surface the audit's instruction layer examines. Quoted here so you can judge it without cloning anything.

---
name: verification-agent
description: Provides a structured verification workflow for code and prompt outputs. Use when asked to validate correctness, catch regressions, or produce confidence-backed signoff.
---

# Verification Agent

## Purpose
Use this skill to verify outcomes with the right level of depth for risk and scope.

## Core Workflow
1. Define the verification target (what must be true).
2. Identify highest-risk failure modes.
3. Select a strategy (quick, targeted, or deep).
4. Execute checks and capture concrete evidence.
5. Report pass/fail status, confidence level, and gaps.

## Minimal Output Contract
- **Status:** pass, fail, or inconclusive.
- **Evidence:** commands, tests, or observable behavior.
- **Risk:** what could still be wrong.
- **Next step:** exact follow-up action if needed.

## Strategy Guide
Use [strategies.md](strategies.md) for practical strategy selection and ready-to-run verification patterns.

SAFEgrade B · trust 89/100 Nothing in the source contradicts what it says it does. Grade A is reserved for packages that have also passed the behavioural sandbox.

What the source does

Filesystem

none-observed

Network

none-observed

Shell

none-observed

Dependencies

pinned

Secrets in source

none-found

Findings (0)

No findings outside the package's declared scope.

Gates applied: no_behavioural_pass.