customer-strategy-forge
Pass
Audited by Gen Agent Trust Hub on Sep 13, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to synthesize artifacts from external customer evidence, which acts as a vector for indirect prompt injection if the source material contains malicious instructions.
- Ingestion points: Untrusted data enters the context through evidence records (interviews, social posts, documents) as defined in
references/evidence-and-gates.md. - Boundary markers: The skill employs a role-based isolation framework (Coordinator, Synthesizer, Critic, Product Mapper) and strict "product-blind" protocols in
references/isolation-and-critic.mdto prevent context leakage. - Capability inventory: The agent is authorized to perform file write operations for persisting artifacts and use search/read tools to audit evidence packs.
- Sanitization: The workflow relies on "redaction review," evidence audits, and an independent critic role to validate claims, though it lacks automated content sanitization for injected instructions within the evidence.
Audit Metadata