The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it reads and processes existing documents from the repository and external policy data provided by the user to generate its output. Ingestion points: The skill reads an existing Platform-Engineering-Constitution.md file and requests the location of existing infrastructure policies to incorporate their intent into the new document. Boundary markers: The instructions do not define delimiters or specific warnings to ignore instructions embedded within the ingested data. Capability inventory: The skill has the capability to write files (Platform-Engineering-Constitution.md) to the repository root. Sanitization: No sanitization or validation logic is defined to prevent embedded instructions in source documents from overriding the agent's behavior during the document generation process.

platform-constitution