legal-spend
Pass
Audited by Gen Agent Trust Hub on Oct 11, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external data from multiple enterprise systems (Accounts Payable, Matter Management, E-billing), creating an inherent attack surface for indirect prompt injection if those source systems contain adversarial content.
- Ingestion points: Data is ingested from Accounts payable system (invoice data), Matter management system (matter codes, descriptions, status), E-billing systems (Brightflag, Legal Tracker, eBillingHub), and Budget systems via MCP as defined in
SKILL.md. - Boundary markers: No explicit instruction-isolation delimiters or 'ignore' warnings for the external data are present in the skill body.
- Capability inventory: The skill body itself does not contain subprocess calls, file-write operations, or direct network requests; it provides logic for the agent to process data provided by external MCP tools.
- Sanitization: The skill does not define specific sanitization or filtering logic for the ingested content, though it mandates review by a licensed attorney before any output-based action is taken.
Audit Metadata