acquisition-strategy
Pass
Audited by Gen Agent Trust Hub on Sep 3, 2026
Risk Level: SAFE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill defines an ingestion point by reading from .agents/growth-context.md at startup. Analysis of the evidence chain confirms the risk is mitigated: 1. Ingestion: Startup reading is scoped to specific fields; 2. Boundary markers: Explicit instructions to surface conflicts and validate compatibility are present; 3. Capability inventory: A full audit confirms zero dangerous capabilities such as subprocess execution, network operations, or file-write authority; 4. Sanitization: Robust rules for excluding PII are established in the acquisition contract reference.
- [DATA_EXFILTRATION]: The skill contains comprehensive privacy guardrails in reference files like acquisition-contract.md and campaign-contract.md, which explicitly mandate the exclusion of sensitive identifiers such as names, emails, and credentials from all analysis and deliverables.
- [COMMAND_EXECUTION]: The skill instructions and associated files were audited for shell commands, dynamic context injection patterns, and remote script execution. No such patterns were identified, and the skill includes an explicit external-action boundary prohibiting unauthorized system interactions.
Audit Metadata