verify-before-done
PassAudited by ClawScan on May 1, 2026.
Overview
This instruction-only skill is coherent and benign, though it may encourage the agent to run relevant checks or inspect available files/logs before saying work is complete.
This skill appears safe and instruction-only. Be aware that it encourages the agent to verify claims using available checks, files, logs, docs, or integrations when relevant, so you should still review any high-impact command or external request before allowing it.
Findings (1)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
The agent may perform lightweight checks or real integration verification instead of relying only on verbal confidence.
The skill may lead an agent to use available tools, commands, or integrations for verification. This is clearly disclosed and directly aligned with the skill's verification purpose.
if code changed, run the most relevant test, build, lint, or minimal execution check ... if an integration changed, make a real request or inspect the real output
Install if you want stricter verification behavior; review any proposed command or real integration request when it could affect external systems or important data.