objection
PassAudited by ClawScan on May 10, 2026.
Overview
This is an instruction-only adversarial review skill with no code or credentials, but it pushes a very strict critical posture and may encourage running or testing artifacts if tools are available.
This appears safe to install as an instruction-only critique skill. Expect harsh, problem-focused reviews, and avoid letting it run code or tests unless you explicitly approve that and can do so in a sandbox.
Findings (2)
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 produce only criticisms and may omit positives, tradeoffs, or reassurance even when those would normally be useful.
The skill deliberately narrows the agent's objective to finding problems and suppresses balanced feedback. This is purpose-aligned for a dissent skill, but users should know it changes the response style strongly.
The goal is maximum problem discovery, not balance. Do NOT self-censor in favor of politeness or efficiency.
Use this skill only when you want adversarial review, and give explicit boundaries if you need brevity, balance, or a limited scope.
If the agent has execution tools, it might try to run user-provided code or tests while reviewing, which could affect the local environment if not sandboxed.
The skill encourages execution or testing as part of verification. That can be appropriate for stress testing, but if applied to code or scripts it should not mean running untrusted artifacts without safeguards.
"This looks correct" | Run it. Test it. Trace the logic step by step.
Require explicit user approval before executing artifacts, run code only in a safe sandbox, and prefer static review when execution is unnecessary or unsafe.