sales-forecast-diagnostic
Pass
Audited by Gen Agent Trust Hub on Oct 26, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill's primary function is to ingest and analyze untrusted external data (CRM records and CSV exports), which creates a potential surface for indirect prompt injection if those records contain hidden instructions.\n
- Ingestion points: The skill ingests data from CRM exports (Salesforce and HubSpot), opportunity snapshots, activity logs, and manual user inputs during the interview phase (SKILL.md).\n
- Boundary markers: While a "Scope boundary" for the project is defined, the instructions do not provide explicit delimiters or "ignore embedded instructions" warnings for the agent when processing external records.\n
- Capability inventory: The workflow explicitly suggests using the agent's code execution "harness" to script reconstructions and reconcile data lists (SKILL.md).\n
- Sanitization: There are no procedures defined for validating, filtering, or escaping the content of external records before they are analyzed or used in the agent's scripting environment.
Audit Metadata