The Agent Skills Directory

[SAFE]: The skill provides explicit instructions for security hygiene, including treating web content as hostile and redacting PII/secrets before sharing context with subagents.- [COMMAND_EXECUTION]: Uses a vendor-specific execution proxy (prefixed with rtk) for terminal actions, such as rtk npm test. This is presented as a standard framework protocol for deterministic execution and monitoring.- [INDIRECT_PROMPT_INJECTION]: The skill acknowledges the ingestion of untrusted external data (via BrowserOS) and provides internal guidelines to mitigate potential risks through context redaction and a 'hostile' trust model for external inputs.

galyarder-specialist