a1-company-profile
Pass
Audited by Gen Agent Trust Hub on Sep 11, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted input in the form of company facts and marketing context files. While this creates a vulnerability surface, it is mitigated by internal logic.
- Ingestion points: User-provided company details and repository context files such as
.agents/marketing-context.md. - Boundary markers: The
source-policy.mdexplicitly requires all claims to trace to allowed sources and separates company statements from unsupported claims. - Capability inventory: The skill is restricted to text generation and does not have file-write or network-access capabilities.
- Sanitization: It includes instructions to replace generic adjectives with evidence and to keep unresolved or unverified facts out of the finished copy.
- [SAFE]: No malicious patterns, obfuscation, or unauthorized access attempts were detected in the skill instructions or supporting files. The logic is focused on content structure and factual validation.
Audit Metadata