account-based-marketing
Pass
Audited by Gen Agent Trust Hub on Sep 1, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill reads from
.agents/product-marketing-context.mdto define its Ideal Customer Profile (ICP) and positioning. If this file is populated with malicious instructions from an external source (e.g., a pull request or untrusted user input), it could lead to the agent following hidden instructions rather than its intended purpose. - Ingestion points: Reads
.agents/product-marketing-context.mdfor core operating context. - Boundary markers: None mentioned; the instructions directly tell the agent to "Read .agents/product-marketing-context.md for ICP definition".
- Capability inventory: The skill performs file writes to
.agents/abm-accounts/[company-slug].mdand.agents/target-account-list.md. - Sanitization: There is no mention of sanitizing the content read from the context file before it is used to generate account briefs or outreach strategies.
Audit Metadata