The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill creates a surface for indirect prompt injection by reading and processing data from external project files that could contain hidden instructions.- Ingestion points: In Step 1, the skill reads .specflow/features/<fid>-<feature-slug>/overview.md and specs.feature. In Step 2, it reads architecture documents D01-D08 and domain-knowledge.md.- Boundary markers: The instructions do not specify the use of delimiters or 'ignore embedded instructions' warnings when reading these files.- Capability inventory: The skill uses the @explore tool to investigate the codebase and has the authority to write new implementation design files to the project directory (Step 10).- Sanitization: There is no mention of escaping, validating, or filtering the content of the ingested files before the agent uses them to generate the implementation design.

203-implementation-design