The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes session logs from AI tools, which may contain attacker-controlled instructions.\n
Ingestion points: The extract_sessions.py script reads session histories from ~/.claude, ~/.codex, and ~/.local/state/opencode.\n
Boundary markers: No specific boundary markers or instructions to ignore embedded content are present in the review workflow.\n
Capability inventory: The skill can execute local scripts, write to a history file, and delete session logs.\n
Sanitization: No sanitization is performed on the extracted conversation content before it is processed by the agent.\n- [COMMAND_EXECUTION]: The skill relies on several local Python scripts to extract data, save reviews, and show trends. All script calls are performed locally based on the intended workflow.\n- [PRIVILEGE_ESCALATION]: The purge_sessions.py script provides a mechanism to delete session files. This is a functional feature for data management, and the skill instructions include checks for user confirmation.\n- [DYNAMIC_EXECUTION]: The save_review.py script uses a static __import__ call to access the sys module for error logging. This implementation is safe as it does not rely on dynamic input for module resolution.

prompt-reviewer