ad-buyer-group-mapper

Pass

Audited by Gen Agent Trust Hub on Sep 13, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process potentially untrusted external data which could contain malicious instructions designed to influence agent behavior.
  • Ingestion points: Data enters the agent context through user-provided evidence including closed-won/lost deal history, buyer interview transcripts, call notes, and review mining as detailed in the 'Clarifying questions' and 'Evidence gate' sections of SKILL.md.
  • Boundary markers: The instructions do not define specific delimiters or explicit safety warnings to the model to ignore instructions embedded within the ingested evidence.
  • Capability inventory: Analysis of the skill body and metadata reveals no use of subprocesses, network tools, file writing capabilities, or remote script execution; the skill's impact is limited to text generation.
  • Sanitization: There is no evidence of input validation, filtering, or escaping logic defined for handling the external data strings.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 13, 2026, 03:32 PM
Security Audit — agent-trust-hub — ad-buyer-group-mapper