The Agent Skills Directory

[SAFE]: The skill consists of methodological guidelines for security review and metadata configuration. No malicious instructions, obfuscation, or unauthorized data access patterns were identified.
[NO_CODE]: No executable scripts, binaries, or shell commands are present within the skill files.
[INDIRECT_PROMPT_INJECTION]: The skill defines a surface for ingesting untrusted application data for review purposes. However, it lacks executable capabilities that could be exploited and explicitly provides safety instructions, such as requiring human approval at action boundaries.

security-review