The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill retrieves a sensitive authentication token from the local file path ~/.config/jetty/token to authorize its operations.\n- [DATA_EXFILTRATION]: The authentication token is sent to the flows-api.jetty.io domain via HTTP headers in curl commands. This domain aligns with the author's ('jettyio') vendor resources and is used for retrieving workflow trajectory data.\n- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by processing untrusted workflow execution data and using it to drive modifications to local files.\n
Ingestion points: Workflow data is ingested from /app/trajectory.json in headless mode and from API responses from flows-api.jetty.io in interactive mode.\n
Boundary markers: No explicit delimiters or instructions are provided to the agent to treat the ingested data as untrusted or to ignore instructions embedded within the JSON content.\n
Capability inventory: The skill utilizes the Edit tool to perform modifications to local runbook files based on the analyzed patterns.\n
Sanitization: No sanitization or validation of the text content within the trajectory data is performed before it is used to generate file edits.

optimize-runbook