competitive-executive-market-brief
Pass
Audited by Gen Agent Trust Hub on Sep 28, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external competitive intelligence reports (LMTY) and internal corporate data (CRM records, call transcripts), creating a surface for indirect prompt injection.
- Ingestion points: Market data and competitor signals are retrieved via tools described in
references/LMTY.mdandreferences/CONTEXT.md. Internal deal, customer, and product data are ingested through connectors described inreferences/INTERNAL.md. - Boundary markers: The skill instructions do not specify explicit boundary markers or unique delimiters to isolate untrusted external content from the prompt instructions.
- Capability inventory: The skill is restricted to information synthesis and report generation; no subprocess execution, dynamic code evaluation, or direct network operations were identified in the skill's scripts or instructions.
- Sanitization: The skill lacks programmatic sanitization of ingested text, relying instead on "epistemic model" guidelines in
references/TRUST.mdthat instruct the agent to distinguish between observed facts and interpretations.
Audit Metadata