The Agent Skills Directory

[SAFE]: No malicious patterns, obfuscation, or unauthorized access vectors were identified within the skill instructions.- [DATA_EXPOSURE]: The skill performs local file read operations on PRD.md and project test files to extract requirement IDs and verify linkages. These operations are essential to its function and do not involve network exfiltration or unauthorized access to sensitive system configurations.- [COMMAND_EXECUTION]: The skill specifies the use of grep for pattern matching within local files. This is a standard, low-risk command utility appropriate for the skill's purpose.- [INDIRECT_PROMPT_INJECTION]: The skill processes potentially untrusted external data (PRDs and source code). Evidence Chain: 1. Ingestion points: PRD.md and test source files. 2. Boundary markers: Absent. 3. Capability inventory: Local file search via grep. 4. Sanitization: Absent. Although the skill lacks explicit delimiters, the logic is constrained to deterministic ID matching, and specific guardrails are included to prevent the agent from synthesizing or 'softening' the audit results based on user pressure or data content.

traceability-audit