Command Claude Code
Scaffold a new SWARM simulation domain — data models, action types, task handler, agents, metrics, tests, and registry wiring — when adding a fundamentally new task category (e.g. medical triage, code review) whose observables or agent actions don't yet exist in swarm/models/ or swarm/agents/. Use /addscenario instead…
Command Claude Code
Add a new metric and wire it into SWARM reporting — use when analysis reveals a measurement gap not covered by existing toxicity, quality-gap, or payoff metrics. Distinct from /addscenario (new parameter config) and /adddomain (new task category with new observables).
Command Claude Code
Scaffold and publish an original blog post to swarm-ai.org/blog/ — handles MkDocs slug generation, metadata headers, nav wiring, and financial-disclaimer enforcement. Use for manually-written essays and research notes; use /evalwriteup instead to auto-generate a post from a Prime Intellect eval run.
Command Claude Code
Scaffold a new YAML scenario config with SWARM conventions — use when testing a new governance configuration or agent motif that uses existing domain infrastructure. Distinct from /adddomain (new task category requiring new data models and handlers) and /addmetric (new measurement on existing data).
Command Claude Code
Fetch review comments on a GitHub PR, apply fixes, push, and reply to each comment thread.
Command Claude Code
Run statistical analysis on SWARM experiment data. Two modes: run a scenario across multiple seeds, or analyze an existing sweep CSV — use when you have run data and need statistical rigor (p-values, confidence intervals, effect sizes). Distinct from /sweep (generates data without analysis), /runscenario (single run…
Command Claude Code
Audit all project metadata files for stale counts, version mismatches, broken references, and missing entries. Reports discrepancies without auto-fixing.
Command Claude Code
Deep codebase audit: spawn category-specific agents, triage findings by severity, fix critical issues.
Command Claude Code
Run the SWARM Track A multi-agent benchmark suite (standardized task conditions, governance presets, behavioral assessments) and auto-compile results to PDF — use when evaluating agent behavior against reproducible baselines before paper submission, or to generate comparison data across model variants. Distinct from…
Command Claude Code
Type-check and build the SWARM visualization game (viz/ TypeScript frontend) in one step — use after editing viz/src/ files, when TypeScript errors need triaging, or before deploying the interactive simulation frontend. Stops on type errors before attempting a build.
Command Claude Code
Update the project version across all files that track it. Takes a version string as argument.
Command Claude Code
Graph-theoretic fleet dispatch — analyze the beads graph with bv, publish ranked per-track digests agents pull from at claim time (push ASSIGN only for strategic beads).
Command Claude Code
Verify that .gitignore won't hide files in a target directory before you create them. Prevents the "wrote 7 files but git can't see them" footgun.
Command Claude Code
Cherry-pick one or more commits onto a new branch from main and open a PR. Useful when a commit landed on the wrong branch or you want to split a multi-commit branch into separate PRs.
Command Claude Code
Atomic work-start. Claim a task before doing any work on it — this refuses if another session already holds it, which is the one thing that physically prevents two sessions from building the same feature (the 2026-07-22 duplicate- 7ge5 incident).
Command Claude Code
Compare results across multiple completed study run directories: compute cross-study descriptive statistics, pairwise tests, and generate a comparison plot.
Command Claude Code
Convert a SWARM paper from markdown to LaTeX (if needed), compile to PDF, and open it.
Command Claude Code
Run a multi-LLM council evaluation on study results. Three expert personas (mechanism designer, statistician, red-teamer) deliberate on the findings using the council protocol.
Command Claude Code
Build and deploy the SWARM blog/docs site to Vercel production.
Command Claude Code
Deploy (or update) the Streamlit interactive demo to Hugging Face Spaces.
Command Claude Code
End-to-end pipeline: parse a Prime Intellect eval run, read the environment source, and auto-generate a blog post from eval metrics. Use --parse-only to just extract structured metrics without writing a post. Distinct from /addpost (manually-written original blog post) — use /evalwriteup when starting from eval…
Command Claude Code
Triage and fix failing GitHub Actions CI runs. Fetches remote failure logs, categorizes errors, reads the relevant source context, and applies fixes — all in one shot.
Command Claude Code
Multi-mode PR operations: create a PR from local changes, resolve merge conflicts on an existing PR, or run quality gates on an external PR. Consolidates the former /reviewexternalpr command (now /fixpr --review).
Command Claude Code
End-to-end research pipeline: sweep parameters, analyze with statistical rigor, generate plots, and scaffold a paper draft. Chains /sweep, /analyzeexperiment, /plot, and /writepaper into a single workflow.
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: