challenge-swarm

challenge-swarm is a command for Claude Code from damusix/atomic-claude. It costs 0 tokens per session (4,587 once invoked), scanned A, original, MIT.

A review command that asks several independent expert perspectives to inspect an existing design, specification, or plan. Its output is a focused challenge report, not a replacement design.

In plain words
What is it for?
Use it to test a plan against concerns such as engineering, data, business, finance, communication, or delivery, then fold the findings back into the design.
Why use it?
Different reviewers can notice different risks, and disagreements between them can reveal assumptions that one review would miss. It helps find problems before implementation begins.

Command for Claude Code

Written for Claude Code: argument-hint in frontmatter. Also seen: mentions CLAUDE.md; mentions subagents; names the AskUserQuestion tool.

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/challenge-swarm
Clone the repo
git clone --depth 1 https://github.com/damusix/atomic-claude

Made for: Claude Code.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for challenge-swarm

README.md
[![agentmods](https://agentmods.dev/badge/commands/damusix/atomic-claude/challenge-swarm.svg)](https://agentmods.dev/commands/damusix/atomic-claude/challenge-swarm)
Your own site
<a href="https://agentmods.dev/commands/damusix/atomic-claude/challenge-swarm"><img src="https://agentmods.dev/badge/commands/damusix/atomic-claude/challenge-swarm.svg" alt="Measured on agentmods" height="20"></a>
Per session 0 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 4,587 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 5.1 $0.00000 $0.04587
Opus 5 $0.00000 $0.02294
Sonnet 5 $0.00000 $0.00917
Haiku 4.5 $0.00000 $0.00459

Measured 6d ago against content hash 76ebc3a82068, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-06, from the pricing page.

Security

Grade A, and why

challenge-swarm 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 6d 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/challenge-swarm.md · 347 lines

How it starts

The opening of the file, as written. The whole thing — 347 lines — stays where its author put it; the contents beside it link to each section on GitHub.

/challenge-swarm

Take an existing design, spec, or plan and subject it to independent expert scrutiny, then map where the experts disagree. Multiple genuinely independent perspectives surface holes any single review pass misses — and the disagreements between perspectives are the highest-value output, not the consensus. Mechanism adapted from Stanford's STORM (perspective-diverse, grounded multi-agent synthesis); the grounding corpus is the artifact's source material — the codebase, when there is one.

This command does not produce a new design. The planning workflow already exists (/atomic-plan). The deliverable is a challenge report the user folds back into the design — surgical findings only, never a rewrite.

Workflow position

hunch → /gather-evidence → /pressure-test → /atomic-plan → /challenge-swarm → /subagent-implementation
        (does it hold?)    (you defend it    (write design   (isolated experts    (build it)
                            in dialogue)      + spec)          attack it)

Parse arguments

$ARGUMENTS may be empty, a topic phrase, or a target document reference. Classify with this precedence (first match wins, single pass):

  1. Token starts with @ → strip the @, treat the remainder as a path. Resolve relative to cwd; absolute allowed. Must exist and end in .mdtarget document. Otherwise print one line: path '<x>' not found (or not markdown) — pass a design/spec path or topic. and continue classifying the rest.
  2. Token ends in .md and exists on disk → target document (same checks).
  3. Anything else → topic phrase. Glob docs/design/*<slug>*.md and docs/spec/*<slug>*.md. One match → use it, confirm in one line. Multiple → numbered list, typed selection. Zero → ask for a path.
  4. Empty $ARGUMENTS → list candidates: design/spec files changed on the current branch (git diff --name-only <base>...HEAD -- docs/design docs/spec), newest first, numbered, typed selection. No candidates → ask: "Which document should the swarm challenge?"

Read the full file on GitHub · 347 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. 6d ago First seen · 347 lines · 0 tokens per session scan A 76ebc3a82068

Subscribe to this mod's changes

challenge-swarm is a command published in the GitHub repository damusix/atomic-claude (84 stars, last pushed today), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 4,587 tokens. 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.