watchlist
Pass
Audited by Gen Agent Trust Hub on Mar 20, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill is configured with the
Bash(python3 *)andGlobtools, enabling the execution of arbitrary Python scripts and file system exploration. - [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection through the processing of external watchlist data.
- Ingestion points: External data is ingested from
watchlist-importand user-provided symbol lists. - Boundary markers: The instructions do not define clear boundaries or 'ignore' directives for processed data.
- Capability inventory: The skill combines data ingestion with powerful capabilities like
Bashand file access (Read/Glob). - Sanitization: While symbol validation is mentioned, there is no evidence of filtering for malicious instructions in the input.
Audit Metadata