tc-tracker
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill manages technical change records containing project metadata, summaries, and session handoff data intended for AI consumption. This creates a vulnerability where malicious instructions stored in one session could influence the behavior of the AI in a future session.\n
- Ingestion points: Text data is ingested via command-line arguments in
scripts/tc_create.pyandscripts/tc_update.py(e.g., fields like--summary,--motivation,--handoff-progress, and--note).\n - Boundary markers: Although the data is stored in structured JSON, there are no explicit delimiters or warnings within the natural language fields to prevent the AI from following instructions embedded in the content.\n
- Capability inventory: The skill is capable of creating and modifying JSON records and directories within the project's
docs/TC/folder.\n - Sanitization: The scripts use standard JSON encoding to escape characters for file storage but do not include sanitization or filtering logic to detect or neutralize natural language prompt injection attempts.
Audit Metadata