The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The workflow involves reading external project configuration which could lead to indirect prompt injection. * Ingestion points: The skill reads state from /REPO_PROFILE.json and analyzes existing codebase patterns to determine authorization structures. * Boundary markers: No specific delimiters or safety instructions (e.g., ignore embedded instructions) are defined for processing these external file inputs. * Capability inventory: The skill has the capability to write and modify backend code, middleware, and test files. * Sanitization: The instructions do not include steps for validating or sanitizing the input data before it influences the generated code or agent behavior.

authz-policy