content-reviewer
Pass
Audited by Gen Agent Trust Hub on Jul 11, 2026
Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection as it is designed to ingest and process untrusted external content (such as influencer posts, links, and captions) to make gate decisions.
- Ingestion points: Untrusted data enters the agent context via the
<content submission or link>argument and content description inputs. - Boundary markers: The instructions do not specify any delimiters or safety guidelines for the agent to treat the reviewed content as untrusted or to ignore any malicious instructions embedded within it.
- Capability inventory: The skill can write review artifacts to
memory/audits/influencer/and update persistent facts inmemory/hot-cache.md. - Sanitization: No sanitization or validation of the input content is defined in the skill logic.
- [EXTERNAL_DOWNLOADS]: The skill contains logic to fetch reference materials, including the auditor-runbook and ART benchmarks, from the author's official GitHub repository (
https://raw.githubusercontent.com/aaron-he-zhu/aaron-marketing-skills/main/) when they are not found in the local file system during standalone installation.
Audit Metadata