meta-accounting-finance-review
Pass
Audited by Gen Agent Trust Hub on Sep 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill consists of instructional markdown and checklists without executable code, scripts, or network operations.
- [INDIRECT_PROMPT_INJECTION]: The skill defines a process for ingesting and analyzing external financial records, business plans, and operating statements, which represents a potential surface for indirect prompt injection. * Ingestion points: Business plans, financial projections, historical financials, and bank/mobile money statements (SKILL.md). * Boundary markers: The skill includes a 'Capability and Permission Boundaries' section that restricts the agent to read-only inspection and prevents unauthorized mutation or publication. * Capability inventory: The skill is limited to analytical tasks, generating reports, and producing evidence registers. It does not spawn subprocesses or perform network writes. * Sanitization: No explicit sanitization or escaping mechanisms are described for the external content being processed.
Audit Metadata