structured-consensus
Pass
Audited by Gen Agent Trust Hub on Sep 24, 2026
Risk Level: SAFENO_CODEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill defines high-level logic for consensus distillation and disagreement mapping. No malicious behavior was identified during analysis.\n- [NO_CODE]: The skill consists exclusively of markdown documentation and YAML schemas. No scripts or executable files are provided.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes external data records, representing a potential ingestion surface. However, the lack of sensitive tool access mitigates the risk.\n
- Ingestion points: Input contract in SKILL.md specifying judgment_records and evidence_records.\n
- Boundary markers: Absent.\n
- Capability inventory: The skill only orchestrates other internal agent skill transitions; no OS-level commands, network calls, or file writes are identified.\n
- Sanitization: None observed.
Audit Metadata