atomic-help

A help router for the atomic-claude workflow. It checks the current Git repository and points you to one suitable command, skill, agent, or program option for what you want to do.

In plain words
What is it for?
Use it without arguments when you are unsure what to do next, with a topic or request for targeted guidance, or with tour for a guided introduction.
Why use it?
It removes the need to remember which tool handles planning, shipping, broken checks, or other common tasks. It gives a focused next step instead of repeating all documentation.

Command

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add commands/damusix/atomic-claude/atomic-help
Clone the repo
git clone --depth 1 https://github.com/damusix/atomic-claude
Per session 57 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 6,048 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00057 $0.06048
Opus 5 $0.00028 $0.03024
Sonnet 5 $0.00011 $0.01210
Haiku 4.5 $0.00006 $0.00605

Measured 3d ago against content hash 6a60ce00c139, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

atomic-help scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 3d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

context/commands/atomic-help.md · 324 lines

How it starts

The opening of the file, as written. The whole thing — 324 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are a routing assistant for the atomic-claude workflow. The user typed /atomic-help because they are unsure what fits their situation. Your job is to classify their state and recommend one next action — not to recite the README.

$ARGUMENTS may be empty, the literal tour (guided walkthrough), a topic keyword (see Step 2.B table), or freeform intent ("I want to ship this", "my CI is broken").

Step 1 — Read git + repo state

Always run these first. They drive routing and the freshness check that gates the tour offer.

git rev-parse --is-inside-work-tree 2>/dev/null
git branch --show-current 2>/dev/null
git status --porcelain 2>/dev/null | head -20
BASE=$(gh repo view --json defaultBranchRef -q .defaultBranchRef.name 2>/dev/null || git config init.defaultBranch || echo main)
git rev-list --count "$BASE"..HEAD 2>/dev/null
git worktree list 2>/dev/null
ls docs/spec/ 2>/dev/null
ls .claude/.scratchpad/ 2>/dev/null
test -f docs/wiki/index.md && echo signals=yes || echo signals=no
test -f CLAUDE.md && echo claudemd=yes || echo claudemd=no

Derive:

  • in_repo — git work tree yes/no
  • branch — current branch name
  • on_base — branch == BASE
  • dirty — any uncommitted changes
  • ahead — commits ahead of base (integer)
  • in_worktree — cwd path includes .claude/worktrees/
  • has_spec — any files in docs/spec/
  • has_scratchpad — any active scratchpad dirs (implies in-flight /subagent-implementation)
  • has_signalsdocs/wiki/index.md present
  • has_claudemdCLAUDE.md present at repo root
  • fresh_repoin_repo AND NOT (has_signals OR has_claudemd OR has_spec) — signals the user has never run the atomic toolchain here

Step 2 — Classify intent

A. No arguments — state-driven recommendation

Pick one primary recommendation from this decision table. Show it first, then ≤2 alternatives. Do not list everything.

State Primary recommendation Why
not in_repo /setup-wiki (after git init) repo not initialized
in_repo + fresh_repo /setup-wiki then /refresh-wiki toolchain never run here
in_repo + on_base + clean /atomic-plan then /subagent-implementation (worktree created at start of loop) start fresh work in isolation
on_base + dirty commit or stash first, then /subagent-implementation (worktree created by the loop) base should stay clean
feature branch + dirty + no spec /atomic-plan (write the contract first) plan before code
feature branch + dirty + has spec /subagent-implementation spec exists, drive the loop
feature branch + has_scratchpad resume /subagent-implementation loop in flight
feature branch + clean + ahead > 0 /review-branch then /commit [push|pr|merge|squash] pre-flight then ship
feature branch + clean + ahead == 0 nothing to ship — back to /atomic-plan or /subagent-implementation empty branch

Read the full file on GitHub · 324 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 3d ago First seen · 324 lines · 57 tokens per session scan A 6a60ce00c139

Subscribe to this mod's changes

atomic-help is a command published in the GitHub repository damusix/atomic-claude (83 stars, last pushed 9d ago), licensed MIT. It adds 57 tokens to every session and 6,048 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.