marketing-plan
Pass
Audited by Gen Agent Trust Hub on Oct 11, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted external data, including client-provided materials (pitch decks, audit outputs, meeting transcripts) and data from various API integrations (Ahrefs, Stripe, GA4). This creates a vulnerability surface where instructions embedded in those external sources could attempt to influence the agent's synthesis of the marketing plan.
- Ingestion points: Untrusted data enters the agent context through the
materials/folder and the research record (research.md) during the INIT phase. - Boundary markers: The skill mitigates this risk by implementing a mandatory 'REVIEW' phase, requiring the user to interactively approve or adjust each of the 13 sections of the plan before it is finalized.
- Capability inventory: The skill possesses capabilities for local file system writes within the
~/marketing-plans/directory and performs network operations through several authorized MCP integrations. - Sanitization: The skill enforces a 'Path safety' rule that slugifies client names and explicitly rejects relative or absolute paths (e.g.,
..,/) to prevent directory traversal attacks.
Audit Metadata