tdd-workflow
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONMETADATA_POISONING
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied requirements and source code to generate tests and implementations. This creates a surface where malicious instructions could be embedded in the data being analyzed.\n
- Ingestion points: Processes user prompts for new features or bug fixes, and reads existing code files during the RED-GREEN-REFACTOR cycle (SKILL.md).\n
- Boundary markers: The skill lacks explicit instructions or delimiters to isolate and ignore potentially malicious instructions embedded in user code or requirements.\n
- Capability inventory: The skill utilizes file system permissions (Read, Edit, Write) and shell execution (Bash) to modify code and run test suites (pytest, npm).\n
- Sanitization: There is no evidence of input validation or sanitization for the code or instructions provided by the user.\n- [METADATA_POISONING]: The skill metadata identifies the author as 'platform-team', which differs from the provided developer identity 'hieutrtr'. This discrepancy, while minor, is a form of deceptive metadata that could imply an official status or origin.
Audit Metadata