competitive-asset-audit
Pass
Audited by Gen Agent Trust Hub on Sep 28, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill implements comprehensive protocols for handling internal and external data. It specifically prioritizes approved company facts and LMTY evidence over content found in the audited assets, ensuring that information is validated before being included in the final report.
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content from competitive assets such as sales decks and websites, which introduces a potential injection surface. However, the methodology explicitly requires the agent to verify all claims against trusted sources and treats the assets as objects of audit rather than sources of instruction, which effectively neutralizes the risk.
- Ingestion points: Competitive assets (files, URLs, folders) retrieved via connected repositories or user uploads as specified in the context acquisition workflow.
- Boundary markers: The Workflow and Methodology files define strict boundaries and classification steps for claims, using a structured audit template.
- Capability inventory: Access to competitive intelligence tools (LMTY), internal context (CRM), and the ability to generate inline reports or save files as requested by the user.
- Sanitization: The skill enforces a Source Hierarchy that outranks external data and requires verification of every material statement using independent evidence.
Audit Metadata