Skill Claude CodeCodex
Phase 2 of an rd-study — compile a decision brief into a ReactiveDynamics model document (.rdj.json), one file per experimental arm, and iterate to zero validation diagnostics. Covers places, transitions, arcs, the three modality axes, structured tokens, expression nodes, and the host.jl registry seam. Use after…
Skill Claude CodeCodex
Phase 3 of an rd-study — prove a validate-clean ReactiveDynamics model is actually plausible before any number is quoted. Runs nine checks (transitions fire, contention exists, the intended resource binds, modality intent, horizon sufficiency, bounded pools, ledger wired, stochasticity) against the brief's…
Skill Claude CodeCodex
Phase 1 of an rd-study — turn a vague business question into a decision brief through Socratic interview. Elicits the decision, the lever, the scarce resource, the process that consumes it, the contention rule, the horizon, and the headline metric; pre-registers what could go wrong. Produces BRIEF.md. Use before…
Skill Claude CodeCodex
Phase 4 of an rd-study — turn simulation output into a decision briefing. Runs the two-arm treatment effect or ensemble summary, checks the effect against the noise, and writes FINDINGS.md to the case-study signature: headline number with error bar, decision rule, why a spreadsheet cannot produce it, collapsible…
Skill Claude CodeCodex
Run a decision study on a resource-constrained, timed process using ReactiveDynamics.jl — answer questions like "what is the fifth scientist worth", "where is the bottleneck", "should we take this deal", "how much capacity do we need". Use when someone needs a defensible number out of a system where something binds…
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: