assess-publication-bias
Pass
Audited by Gen Agent Trust Hub on Sep 24, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: The skill consists exclusively of markdown instructions and does not include any scripts, executable commands, or network operations.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external evidence records and research logs. While this creates a theoretical surface for indirect prompt injection, the skill lacks any capabilities (such as shell execution or file system writes) that could be exploited by malicious data.
- Ingestion points: Inputs defined in SKILL.md (evidence_records, registry_records, search_log).
- Boundary markers: None provided in the instructions.
- Capability inventory: No scripts, tools, or subprocess calls are defined in this skill.
- Sanitization: None specified for input data.
Audit Metadata