ideal-customer-profile
Pass
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: No malicious patterns, obfuscation, or unauthorized access attempts were detected in the skill instructions or evaluation suite.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external research data (surveys, interviews, transcripts), which is an inherent attack surface for indirect prompt injection.
- Ingestion points: Step 1 involves gathering research data from various external sources.
- Boundary markers: The skill uses a draft/confirmation workflow where synthesized content is shown to the user before being committed.
- Capability inventory: The skill can read from and append to local markdown files (
/foundation/brain.mdand/context/skill-sessions.md). - Sanitization: The skill relies on human-in-the-loop verification (Step 3) to review and approve all proposed updates to the core 'brain' file.
Audit Metadata