{"id":16863,"plugin_id":"plugins_6a6a29948a7c8191ace6787d5ae074bb","kind":"skill","collection_source":null,"comparison_source":null,"observed_at":"2026-09-30T23:13:50.547Z","digest":"84d75f35a43b7392eadbbf05a8383f573a80bdbdaf2b41074b3a261147a57bb3","against":null,"payload":{"description":"Execute a CMA-ES calibration from confirmed Jinkō inputs: assemble the model, protocol, output sets, fitness data tables, parameter priors, and optimizer options; create and run the Calibration; and return the supported results. Use when the user wants to perform a CMA-ES calibration, not when they need to choose a calibration strategy, infer priors, design objectives, or decide whether results are acceptable.","included_files":[{"relative_path":"agents/openai.yaml","size_in_bytes":260}],"name":"jinko-task-cmaes","skill_md_contents":"---\nname: jinko-task-cmaes\ndescription: >-\n  Execute a CMA-ES calibration from confirmed Jinkō inputs: assemble the model,\n  protocol, output sets, fitness data tables, parameter priors, and optimizer\n  options; create and run the Calibration; and return the supported results.\n  Use when the user wants to perform a CMA-ES calibration, not when they need\n  to choose a calibration strategy, infer priors, design objectives, or decide\n  whether results are acceptable.\ncompatibility: >-\n  Check set-up with jinko-sdk-setup. Creating and running calibrations requires\n  write and run permissions.\nmetadata:\n  author: Nova In Silico\n  requires_sdk: \">=1.8,<2.0\"\nlicense: MIT\n---\n\n# CMA-ES Calibration Task\n\nExecute a confirmed calibration specification. Do not invent objectives,\nconstraints, parameter priors, optimizer options, or acceptance criteria.\n\n> **PREREQUISITE:** This skill needs an initialized `jinko-sdk` connection and an\n> SDK satisfying its `metadata.requires_sdk` range. Run the `jinko-sdk-setup` skill\n> (`../jinko-sdk-setup/SKILL.md`) and proceed only once its check passes. If that\n> skill is not found, install it from `novainsilico/jinko-skills`.\n\n## Inputs\n\nRequire:\n\n- a model SID;\n- parameter priors with physical bounds;\n- `seed`, `thresholdWeightedScore`, `numberOfIterations`, and `populationSize`;\n- at least one fitness source: calibration-ready data tables and/or an advanced\n  output set containing objectives;\n- any protocol, simple output set, advanced output set, folder, and name needed\n  by the specification.\n\nIf quantitative evidence has not yet been converted into a calibration-ready\ntable, use `jinko-task-extract-data-table`. Use `jinko-data-table`,\n`jinko-output-set`, `jinko-model`, and `jinko-protocol` only for their respective\nJinkō object mechanics.\n\n## Workflow\n\n1. Resolve every input to its intended SID and snapshot. Present missing or\n   ambiguous inputs instead of guessing.\n2. Use `jinko-calibration-cmaes` and its bundled creation script. Review its\n   dry-run output before applying it. The script owns parameter encoding,\n   fitness-table eligibility, bound scaling, creation, and post-creation sanity;\n   stop on an error and surface warnings.\n3. Return the created calibration SID, revision, snapshot, URL, and effective\n   options for confirmation.\n4. Use the lower-level run script to perform pre-launch sanity, launch, and wait\n   for a terminal state. Do not relaunch a terminal snapshot; create or update a\n   configuration so the intended change has a new snapshot.\n5. Use the lower-level inspection interfaces to collect the final status,\n   stopping reason, performance, results summary, objective weights, and the\n   patient sorted first by `optimizationWeightedScore` when available. Fetch\n   per-patient scalars, timeseries, errors, or augmented data tables only when\n   their required selectors are present in the result metadata.\n\n## Return\n\nReturn:\n\n- calibration SID, revision, snapshot, and URL;\n- effective input references, priors, and optimizer options;\n- sanity warnings, terminal status, stopping reason, and performance;\n- supported result payloads and best-patient identity, with the iteration and\n  scenario arm needed for subsequent result calls;\n- a concise account of unavailable requested outputs.\n\nDo not claim a separate run ID, convergence analysis, score-evolution curve,\nbest-patient parameter values, parameter posterior, or simulation-vs-data plot\nunless the returned API payloads actually provide the required data.\n"},"changes":[],"summary":"First saved snapshot. No earlier version is available for comparison.","summary_kind":"deterministic","summary_metadata":{}}