The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze external content including project requirements, design schemes, and source code. This creates a surface where malicious instructions embedded in the analyzed data could attempt to influence the agent's behavior.
Ingestion points: The skill reads external requirements, design schemes, code diffs, and specifications as described in the 'Design checks' and 'Code checks' sections of SKILL.md.
Boundary markers: There are no explicit delimiters or markers defined to isolate untrusted external data from the agent's instructions.
Capability inventory: The skill is instructed to execute command-line tools including git diff and project-specific testing, linting, and build scripts as noted in SKILL.md.
Sanitization: The instructions do not include steps to sanitize, escape, or validate external content before it is processed by the model.

dl-review