ai-risk-planning
Pass
Audited by Gen Agent Trust Hub on Jun 22, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill invokes a local Python script (scripts/validate_risk_plan.py) during the initial 'Step 0: Data Confirmation and Verification' phase. This script is part of the skill package and is used solely for validating the schema and integrity of input fields like company names and loan amounts. It does not perform network operations or access sensitive system files.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted external inputs, including company reports, due diligence files, and annual reports. This ingestion creates a surface for indirect prompt injection. However, the skill mitigates this through strict 'anti-lazy' instructions (Workflow Steps 0, 2, and 3), a mandatory validation script, and clear boundary definitions in the task planning logic.
- Ingestion points: 'Optional Input Files' (company_info, due_diligence_report, annual_report) defined in SKILL.md.
- Boundary markers: Not explicitly defined for the contents of the processed files.
- Capability inventory: Execution of a local validation script (scripts/validate_risk_plan.py) as defined in SKILL.md.
- Sanitization: Step 0 enforces input validation and consistency checks before any risk planning logic is executed.
Audit Metadata