The Agent Skills Directory

[COMMAND_EXECUTION]: The skill interacts with the local opencode command-line utility in scripts/agent_sessions/opencode.py to fetch session listings and database paths. These calls are performed using subprocess.run with hardcoded arguments and binary path resolution via shutil.which. The execution is restricted to the specific local tool and does not utilize a shell, mitigating typical injection risks.
[DATA_EXPOSURE]: The skill performs broad scanning of local application data directories (e.g., ~/.claude, ~/.codex, ~/.opencode, ~/.aside) and SQLite databases to reconstruct conversation histories. This access is the primary functionality of the skill, and the extracted data is returned to the agent context for processing without any external network exfiltration observed.

coding-agent-sessions