transcript-cleanup
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted transcript data which could contain malicious instructions.
- Ingestion points: Transcript content is read from local files like raw.md during analysis and review phases.
- Boundary markers: The instructions do not specify the use of delimiters to separate untrusted transcript text from agent prompts.
- Capability inventory: The agent can execute local Python scripts and perform web searches based on transcript content.
- Sanitization: No sanitization is performed on transcript inputs.
- [COMMAND_EXECUTION]: The skill utilizes local Python scripts (check_cleaned.py, glossary_candidates.py, locate.py) to perform analysis and validation. These scripts are executed by the agent and in the provided test suite.
Audit Metadata