The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill architecture facilitates the ingestion and processing of data from external, untrusted sources via the Matrix protocol.
Ingestion points: The matrix_worker_task and spawn_timeline_subscriber functions in SKILL.md are designed to receive real-time events and timeline updates from Matrix homeservers and room participants.
Boundary markers: The provided implementation code does not include specific boundary markers or instructions to isolate or ignore potentially malicious commands embedded within ingested message content.
Capability inventory: The skill utilizes the matrix-sdk which requires network access to communicate with remote homeservers and file system access for persistent storage via the sqlite_store in the build_client function found in SKILL.md.
Sanitization: The instruction set lacks specific logic for sanitizing, escaping, or validating the content of incoming Matrix events before they are processed or rendered in the UI context.

robius-matrix-integration