fin-guru-research
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from external web sources, which creates an entry point for instructions embedded in third-party content to influence the agent.
- Ingestion points: External data is collected via
exa,bright-data, and fallback tools likeWebSearchandWebFetchas described in the 'Capability probe' section ofSKILL.md. - Boundary markers: The skill attempts to mitigate injection by requiring citations to be wrapped in
START/ENDtags as defined in the 'Requirements' section. - Capability inventory: The skill executes several local Python CLI tools (
screener_cli.py,moving_averages_cli.py,momentum_cli.py,volatility_cli.py,data_validator_cli.py,itc_risk_cli.py) and performs data synthesis/handoff tasks. - Sanitization: The workflow includes 'Source Validation' and 'Data integrity verification' steps, utilizing
data_validator_cli.pyto check for 100% quality before analysis.
Audit Metadata