finance-based-pricing-advisor
Pass
Audited by Gen Agent Trust Hub on Aug 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill framework processes untrusted user data in the form of financial metrics that are interpolated into the assessment logic without explicit sanitization or boundary markers.
- Ingestion points: Data is gathered from the user in 'Step 0: Gather Context' (ARPU, churn, pricing tiers) and 'Step 3: Evaluate Current State' (MRR, CAC, LTV) in
SKILL.md. - Boundary markers: The instructions lack specific delimiters or system-level directives to ignore instructions that might be embedded within the user-provided financial data.
- Capability inventory: The skill utilizes an 'ExecutionProxy' for terminal commands (
rtkprefix), interacts with an 'IssueTracker' (Linear), and maintains 'Durable Memory' via file writes to Obsidian, which represents a capability set that could be misused if an indirect injection were successful. - Sanitization: No explicit input validation or escaping mechanisms are described to handle potentially adversarial content in the financial metrics.
Audit Metadata