The Agent Skills Directory

[SAFE]: The skill is composed entirely of natural language instructions and does not include any executable scripts, binaries, or software dependencies.\n- [SAFE]: No malicious prompt injection patterns or instructions to bypass safety protocols were detected. The skill includes guardrails that explicitly prevent the agent from performing code reviews or modifying files.\n- [SAFE]: The skill does not contain any network operations, data exfiltration logic, or hardcoded credentials. Its file access is restricted to local markdown documents for analysis purposes.\n- [SAFE]: There is no evidence of obfuscation, hidden URLs, or dynamic execution patterns. While the skill ingests untrusted markdown files, it lacks the dangerous capabilities required for a high-risk indirect prompt injection attack.

review-spec