The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses dynamic context injection via the !command`` syntax to execute local binaries (ops-infra, ops-ci, and ops-external) from the plugin's bin directory at load time to populate the dashboard data.
[COMMAND_EXECUTION]: Orchestrates various CLI tools including aws, gh, and sentry-cli through the Bash and Monitor tools to investigate service health and stream logs.
[PROMPT_INJECTION]: The skill presents an indirect prompt injection surface by processing external data from Sentry error logs and project health checks. Evidence chain: 1. Ingestion points: Sentry issue descriptions and ops-external output; 2. Boundary markers: Absent; 3. Capability inventory: Bash, Agent, and TaskUpdate tools; 4. Sanitization: Absent. Risk is mitigated by the use of AskUserQuestion to confirm agent dispatch and fixes.

ops-fires