The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill analyses user-provided file content, which serves as an ingestion surface for indirect prompt injection. This is a functional requirement of its quality assurance role.
Ingestion points: SKILL.md and files in the 'references/' directory are read for analysis as specified in the evaluation workflow.
Boundary markers: Absent; there are no explicit delimiters defined to isolate the audited content from the skill's auditing logic.
Capability inventory: Capabilities are restricted to reading files and generating text reports; no network access, shell execution, or file system modification capabilities are present.
Sanitization: Absent; input is processed directly as text for evaluation without specific sanitization routines.

skill-trace-checker