foundations-ai-planning-search
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is a pedagogical and practical resource for implementing planning algorithms. It correctly identifies and documents common pitfalls in agentic planning, such as conflating executable sequences with safe ones.
- [SAFE]: The provided Python script
scripts/replay_plan.pyis a deterministic state-transition validator. It performs schema validation on its input and only conducts set operations and arithmetic in memory. It lacks network access, shell execution, or sensitive file system operations. - [SAFE]: All external references in
data/sources.jsonand the documentation point to established academic institutions (e.g., UC Berkeley), well-known technology companies (e.g., Google), and standard preprint servers (arXiv). These are verified as legitimate resources for the field of AI planning. - [SAFE]: There is no evidence of prompt injection, code obfuscation, or persistence mechanisms. The skill maintains a high standard of transparency and theoretical grounding.
Audit Metadata