lead-researcher
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill acts as an orchestrator for various research-related sub-skills. It follows a structured, multi-stage pipeline (Intake, Hypothesis, Literature, Experiment, etc.) that requires explicit user confirmation at critical decision gates.
- [SAFE]: The local execution path (Stage 7c) uses the
ollama-multi-model-writersub-skill to run tasks locally on the user's GPU using established models (DeepSeek-R1, Phi-4, GLM-4). It provides clear instructions for users to pull these models via official Ollama commands, representing standard local development practices. - [SAFE]: There are no signs of prompt injection, data exfiltration, or obfuscation. The skill explicitly includes a 'No Fabrication' principle, instructing the agent never to generate fake data, citations, or results.
- [SAFE]: All metadata and instructional content align with the stated purpose of a lead researcher tool, and no suspicious command execution or privilege escalation patterns were detected.
Audit Metadata