The Agent Skills Directory

[SAFE]: The skill does not contain any malicious patterns such as prompt injection, obfuscation, or persistence mechanisms.
[DATA_EXPOSURE]: The skill accesses and reads other skill files (e.g., SKILL.md and reference files) within the workspace to perform its audit. This is the primary intended function of the skill and does not involve accessing sensitive system directories (like .ssh or .aws) or exfiltrating data to external servers.
[REMOTE_CODE_EXECUTION]: There are no commands or instructions that download or execute remote code, install packages, or perform dynamic execution.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from other files (the skills being reviewed). While this is an attack surface, the skill's functionality is limited to generating a textual report and does not include risky capabilities like file writing or network access that could be abused by malicious content in the reviewed files.

skill-review