القائمة
Scientific reasoning via Jnana CoScientist — hypothesis generation, evaluation, and parameter bounding
التثبيت باستخدام Codex أو Claude انسخ هذا Prompt والصقه في Codex أو Claude أو مساعد آخر ليراجع صفحة Skill ويثبّتها لك.
استنادا إلى تصنيف SOC المهني
| name | jnana-reasoning |
| description | Scientific reasoning via Jnana CoScientist — hypothesis generation, evaluation, and parameter bounding |
| metadata | {"openclaw":{"requires":{"env":["OPENAI_API_KEY"],"bins":["python3"],"anyBins":["python3.11","python3.12"]},"primaryEnv":"OPENAI_API_KEY"},"dependencies":["jnana","academy"]} |
Access Jnana's CoScientist for hypothesis-driven scientific reasoning. This is Layer 2 of the 4-layer hybrid architecture.
The reasoning bridge uses a two-tier prompting strategy:
Tier 1 — Recommendation: recommend_next_action() evaluates the current state
and recommends the next task type (computational_design, molecular_dynamics, analysis,
free_energy, or stop).
Tier 2 — Parameter Bounding: bound_parameters() takes the recommendation and
generates a bounded parameter configuration for the selected skill.
jnana.set_research_goal — Initialize reasoning for a research goaljnana.recommend_next_action — Get next recommended task type (Tier 1)jnana.bound_parameters — Get parameter config for a skill (Tier 2)jnana.evaluate_results — Evaluate artifacts against hypothesesjnana.check_convergence — Check if research goal is metpython skills/jnana-reasoning/scripts/reason.py --help
python skills/jnana-reasoning/scripts/reason.py set-goal "Design a binder for target X"
python skills/jnana-reasoning/scripts/reason.py recommend --previous-run starting
python skills/jnana-reasoning/scripts/reason.py bound-params --skill bindcraft --task-type computational_design
python skills/jnana-reasoning/scripts/reason.py evaluate --artifact-ids abc123 def456
python skills/jnana-reasoning/scripts/reason.py check-convergence
action: Reasoning action — "set_goal", "recommend", "bound_params", "evaluate_results", "check_convergence"research_goal: Research goal description (for "set_goal")previous_run_type: Previous run type for context (for "recommend")skill_name: Target skill for parameter recommendations (for "bound_params")task_type: Task type for parameter schema (for "bound_params")artifact_ids: Artifact IDs to evaluate (for "evaluate_results")Search and select enzyme sequence homologs from a database. Use when users provide a query sequence and want to discover sequence homologs of the query sequence in a database.
Molecular dynamics simulations using OpenMM with MDAgent-style automation and free energy calculations via MM-PBSA
Convert scientific papers into executable computational workflows and MCP tools
Protein language model embeddings, diversity sampling, and mutation prediction using ESM or GenSLM
Protein structure prediction using Chai-1 or AlphaFold with confidence scoring, quality assessment, and critic evaluation
Analysis of molecular dynamics trajectories (RMSD, RMSF, contacts, hotspot analysis)