The Agent Skills Directory

[SAFE]: No malicious patterns, obfuscated code, or unauthorized network operations were detected. The skill's behavior is consistent with its stated purpose of codebase maintenance.
[COMMAND_EXECUTION]: The skill uses local shell commands like find, grep, and git to analyze codebase statistics and identify stale or redundant files. These operations are diagnostic and do not attempt to modify system settings or escalate privileges.
[EXTERNAL_DOWNLOADS]: The skill references standard development tools including vulture, knip, and autoflake as optional dependencies for advanced analysis. It does not attempt to automatically download or install these from untrusted sources; instead, it checks for their presence on the user's system.
[PROMPT_INJECTION]: The skill reads and processes user-controlled files (Python scripts, JS, Markdown) to detect bloat patterns. While this creates a surface for indirect prompt injection, the risk is mitigated by explicit instructions requiring user review and manual approval for all proposed remediation actions (e.g., code deletion).

bloat-detector