{"data":{"kind":"file","path":"README.md","version_id":"bo8qcy101h0ibpv8bcbwg6fj","entry":{"name":"README.md","path":"README.md","is_directory":false,"size":972,"modified_at":"2026-05-11T14:32:06.286000","content_hash":"628e36102a5fc8d86c20cdef1ae07cbdbc5b3b2fbbdfa261216926973f50c404"},"entries":[],"content":"# science-sim-materials\n\nMaterials Science Sim field environment for in-silico candidate ranking and simulation planning.\n\n## Tool Loop\n\n```text\ninspect_problem(target_id)\nlist_candidate_materials()\nrank_materials()\nsimulate_material(material_id)\ncompare_materials(material_ids)\nsubmit_decision(material_id, decision, expected_performance, expected_stability, synthesis_risk)\n```\n\n`rank_materials()` is a noisy screening tool. `simulate_material()` exposes measured values used for the final decision.\n\nDecision values:\n\n- `select_for_application` when the measured material clears performance, stability, synthesis-risk, and budget constraints.\n- `select_as_research_lead` for high-option-value active-learning candidates that are not necessarily deployment-ready.\n- `reject` when measured evidence rules out the available candidates.\n\nThe reward includes a hard-constraint score so deployment selections that violate the application envelope cannot receive full credit.\n","encoding":"utf-8","truncated":false,"total_bytes":972},"status":null}