The Agent Skills Directory

[Command Execution with External Arguments]: The skill utilizes bash and jq to reconstruct session trajectories. User-supplied arguments like workspace_path and session_id are incorporated into shell scripts. Without rigorous validation, this could present a potential surface for path traversal or unintended command execution if malicious inputs are provided to the agent.
[Sensitive Directory Access]: The auditing process explicitly targets session logs stored in hidden user directories, specifically ~/.claude, ~/.pi, and ~/.codex. These locations contain historical records of AI interactions, which may include sensitive code, configuration data, or previous prompts and responses.
[Indirect Prompt Injection Surface]: The skill is designed to ingest and process untrusted data in the form of session JSONL logs. This creates an attack surface where a maliciously crafted log file from a previous run could attempt to influence the agent's auditing logic or result reporting during the reconstruction process.
[File System Discovery Operations]: The procedure involves using find and ls to discover session files across various host directories. While necessary for its intended purpose as an audit tool, this requires the agent to operate with broad read permissions over potentially sensitive areas of the file system.

investigate-run