Code graphs, wikis, and research-backed agentic coding. Deterministic tools for nondeterministic workflows, in one opinionated Claude Code config.
This repository also configures its own agents. See what atomic-claude tells them →
Command Claude Code
Route a lost user to the right atomic verb, skill, agent, or binary subcommand. Bare invocation reads git state and recommends one next step. Topic keyword or freeform intent → focused pointer. tour → guided walkthrough of the system. Help router, not duplicated docs.
Command Claude Code
Write a design doc (concepts, business rules, approaches) and a checkpoint-table spec (contract) for non-trivial work; inline spec only for trivial. Gauges triviality; loops spec authoring with subagents. Human-facing artifact, Mermaid diagrams allowed.
Command Claude Code
Autonomous delivery: plan, run the implement→review loop, ship. Takes a task or a GitHub issue number and an optional merge verb. Asks one question, how to merge, and only when the verb was not given.
Command Claude Code
Multi-lens design challenger. Reads a design/spec/plan/proposal, profiles it (who can reach it, whose data, what being wrong costs), then seats 3-6 expert lenses from a 30-lens catalog spanning engineering, data/ML, business, finance, communication, and delivery — each seated only with a cited stake, reflexive picks…
Command Claude Code
Stage, commit, and optionally ship further. Pass an escalation token (push, pr, merge, squash, squash merge) to skip the prompt. With no token, commits then asks how far to ship. Delegates message format to the atomic-git-discipline skill.
Command Claude Code
Bootstrap and maintain project documentation surfaces. Two modes: bootstrap (discover doc files, index them in CLAUDE.md) and authoring (scan for unindexed docs, match diff against indexed surfaces, walk stale/incomplete/missing items with Yes/Later/Remind/Skip).
Command Claude Code
Review and act on pending reminders. Bare invocation shows all reminders as an indexed list; cron-fired invocation (/follow-up due ) surfaces the specific reminder and waits for a response; /follow-up review lists stale project follow-up entries for per-item disposition (extend/close/promote/skip). Transport-aware…
Command Claude Code
Pre-design evidence gathering. Takes a hunch or hypothesis and chases it through primary sources (context7, official docs, source code, ast-grep, run-it experiments) before any spec is written. Returns SUPPORTED / UNSUPPORTED / MIXED / INCONCLUSIVE with cited evidence trail. Pairs with /pressure-test and precedes…
Command Claude Code
Scan and clean up stale git state — worktrees, local branches, optionally remote tracking refs. Dispatches a read-only scan via atomic prompt git-cleanup, presents an indexed report, asks user which to clean. No destructive ops without explicit confirmation. Defaults: 30-day staleness, local-only.
Command Claude Code
Socratic challenger for design decisions. Pressure-tests assumptions, surfaces contradictions, forces fuzzy maybes into yes/no — through questions only, never producing code or artifacts. Pairs with /atomic-plan as a pre-approval gate.
Command Claude Code
Implement→review subagent loop with no planning phase, for a fix whose cause is known and whose shape is obvious, however many files it touches. No spec, no worktree, no docs or signals at the end; one audit. Routes to /subagent-diagnose on an unknown cause and to /atomic-plan on an open approach or contract choice.
Command Claude Code
Refresh the project wiki. Handles both repo scope (docs/wiki/ inside the current repo) and realm scope (a cross-repo wiki at /wiki/). Scope auto-detected from block or directory structure. Dispatches atomic-wiki-inferrer, which reads the matching pipeline reference and executes it.
Command Claude Code
Schedule a reminder. Creates a reminder file and schedules it via cron ( = 1h). Degrades silently to file-only when scheduling tools are unavailable.
Command Claude Code
Open an issue against the atomic-claude repo itself (bugs/feature requests with the installed config, not the user's current project).
Command Claude Code
Open a atomic GitHub issue via gh. Bug report or feature request — auto-detect from user's description.
Command Claude Code
Session retrospective. Mines session history extracted by atomic retro extract and this conversation for friction, corrections, and misbehavior; cross-references installed artifacts; walks findings one at a time. Persists a run log so later runs detect drift. Use after long sessions or repeated friction.
Command Claude Code
Review the current branch's diff against base by dispatching atomic-reviewer. No orchestration loop, no spec required — pre-flight before /commit pr or /commit merge.
Command Claude Code
Capture what changed this session and why, scoped to the current branch. Read by ship verbs when synthesizing the commit message; deleted after a successful commit.
Command Claude Code
Bootstrap the current repo for atomic-claude use. Audits .gitignore entries, docs/ layout, and presence of CLAUDE.md. Proposes only what's missing — never overwrites. No commits.
Command Claude Code
Multi-agent failure-investigation orchestrator. Parallel to /subagent-implementation — same scratchpad + investigator + builder/surgeon + reviewer + FOLLOWUPS pattern — but starts from a failure, not a spec. Two modes share one loop: ci (failed CI run is the brief seed) and bug (freeform symptom paragraph is the brief…
Command Claude Code
Orchestrate the implement→review subagent loop from an approved spec. Fresh-context implementer and reviewer per checkpoint, commit per green iteration, then verify, document, audit, and refresh signals over the task's range.
Command Claude Code
Soft-undo the last commit. Restores its changes as staged, leaves working tree intact. Refuses if HEAD is already pushed to a tracked remote.
Command Claude Code
Spawn a background Haiku-backed subagent to watch CI for the current branch (or specified target). Provider-agnostic — the subagent inspects project signals to identify the CI system (GitHub Actions, GitLab CI, CircleCI, etc.) and picks the right CLI. Returns immediately; reports back when CI reaches a terminal state.
Command Claude Code
Audit a standing codebase for accumulated slop — comment noise, AI-tell doc prose, speculative abstraction, reinvented stdlib, duplicate helpers, dead code, swallowed errors, convention drift. Fans out read-only auditors sharded by wiki domain and writes an indexed report into a scratchpad for you to read. /deslop…
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: