customer-panel-of-experts
Pass
Audited by Gen Agent Trust Hub on Jun 12, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill processes persona data originating from external tool scans (stored in the 'personas/' directory and 'icp-profile.md'). This ingestion of external data constitutes a surface for indirect prompt injection, as malicious instructions within the source data could attempt to influence the agent's behavior.
- Ingestion points: SKILL.md (Step 0) references the 'personas/' directory and 'icp-profile.md' as the primary data sources for building the expert panel.
- Boundary markers: No specific delimiters or 'ignore' warnings for the raw data interpolation are prescribed; however, the skill mandates a structured synthesis format for the final output.
- Capability inventory: The skill has access to Bash, Write, and Agent tools, which provides a significant capability surface if the agent were to obey instructions embedded in the ingested data.
- Sanitization: The skill explicitly mandates that the agent scrub PII (names, emails, account IDs) and quotes from the data before output, which serves as a mitigation against data exposure.
Audit Metadata