The Agent Skills Directory

[PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection through the processing of external network data.\n
Ingestion points: Reads and interprets diagnostic output from avahi-browse (mDNS discovery), ot-ctl (Thread network state), and curl responses from the local OpenThread Border Router API.\n
Boundary markers: No specific delimiters or safety instructions are provided to the agent to distinguish between tool results and potential embedded malicious instructions.\n
Capability inventory: The skill has access to Bash, Edit, and Write tools, which could be used to execute commands or modify local files based on injected data.\n
Sanitization: There is no evidence of sanitization or validation logic to filter external tool outputs before processing.\n- [COMMAND_EXECUTION]: Provides a wide array of command-line instructions for managing smart home devices using chip-tool, ot-ctl, and docker. It includes a Docker Compose configuration that requires privileged: true to access hardware radios, representing a necessary but elevated privilege level for the specified task.\n- [EXTERNAL_DOWNLOADS]: Fetches source code and build dependencies from the official GitHub repositories for the Matter SDK (Project Chip) and OpenThread. These are standard, well-known resources for smart home development.

matter-thread