The Agent Skills Directory

[SAFE]: The skill's primary purpose is configuration management for the kickoff tool, involving reading and writing model mappings to a local kickoff.yaml file.- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from existing kickoff.yaml files (Step 3), which serves as an ingestion point for potentially untrusted data. While it lacks explicit boundary markers, it includes a validation step (Step 5) that verifies all entries against the system's detected harness capabilities. Its capabilities are limited to writing configuration files and suggesting user-initiated skill creation (Step 7).

kickoff-setup