terminology-management
Pass
Audited by Gen Agent Trust Hub on Sep 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data which could potentially influence agent behavior or script execution.
- Ingestion points: The skill reads from a source markdown corpus and a local
glossary.jsonfile to identify and manage terms (SKILL.md). - Boundary markers: No explicit delimiters or instructions to ignore embedded instructions within the processed corpus are provided to the agent.
- Capability inventory: The skill executes several local Python scripts (
validate_glossary.py,term_generate.py,term_cal_batch.py,term_edit.py, andterm_read.py) usinguv run(SKILL.md). - Sanitization: Although the skill performs schema validation via
validate_glossary.py, there is no evidence of sanitization for the natural language content extracted from the corpus.
Audit Metadata