The Agent Skills Directory

[PROMPT_INJECTION]: The skill presents a surface for indirect prompt injection by design, as it instructs the agent to analyze local repository files for discovery of patterns. Ingestion points: Target repository files such as docs/schema-reference.md, keys.ts, and existing repository implementations as specified in the discovery instructions. Boundary markers: Absent; there are no instructions to delineate untrusted code or documentation from the agent's task instructions. Capability inventory: The skill produces data models, schema documentation updates, and repository-layer code suggestions. Sanitization: Absent; the skill does not specify any filtering or sanitization of the content extracted from the local codebase.

dynamodb-design