The Agent Skills Directory

[SAFE]: No security issues detected. The skill implements Matrix communication using the standard matrix-nio library and follows secure development practices.
[COMMAND_EXECUTION]: The matrix-doctor.py script uses subprocess.run to check for and install the required matrix-nio[e2e] package. This is a legitimate setup procedure and is hardcoded to install a specific, well-known library.
[DATA_EXPOSURE]: Matrix credentials (access tokens and device keys) are stored in the user's home directory (~/.local/share/matrix-skill/store/) with restricted file permissions (0o600), preventing unauthorized access by other users on the system.
[PROMPT_INJECTION]: The skill processes message content from Matrix rooms, which constitutes an ingestion point for untrusted data (Indirect Prompt Injection surface). However, this is inherent to a chat communication skill.
Ingestion points: matrix-read-e2ee.py and matrix-read.py read message bodies from the Matrix API.
Boundary markers: None present in the script output; the agent is expected to handle the message content as data.
Capability inventory: The skill has capabilities for network requests (via Matrix API), file reading/writing, and command execution (via python3 and uv as defined in allowed-tools).
Sanitization: Filenames in matrix-download-e2ee.py are sanitized using Path.name to prevent path traversal.

matrix-communication