The Agent Skills Directory

[COMMAND_EXECUTION]: The skill mandates a voice notification via curl to localhost:31337 whenever a workflow (Scaffold, Interview, CheckCompleteness, Reconcile, Seed, Append) is invoked. Additionally, the Seed and CheckCompleteness workflows utilize git log to extract project history and metadata from the local repository.\n- [PROMPT_INJECTION]: The Seed workflow has a potential surface for indirect prompt injection because it ingests data from external, potentially untrusted project files to bootstrap new artifacts.\n
Ingestion points: Workflows/Seed.md reads repository files including README.md, package.json, tsconfig.json, PRD.md, SPEC.md, acceptance.yaml, and various test files (*.test.ts).\n
Boundary markers: The skill does not provide explicit instructions or delimiters to the agent to disregard instructions that might be embedded in the ingested repository files.\n
Capability inventory: The skill possesses the capability to read and write to the file system and execute shell commands (curl and git).\n
Sanitization: There is no evidence of content sanitization, filtering, or validation for the data ingested during the Seed process.

ISA