work

A command that coordinates several language-model agents through planning, coding, testing, review, and pull-request workflows. It supports feature work and major or minor bug fixes with different levels of review.

In plain words
What is it for?
Use it to plan features, implement changes, fix major bugs using test-driven development (writing a failing test before the fix), or quickly fix minor bugs and commit them.
Why use it?
It organizes the repeated steps of software work and brings in multiple reviewers where required. This helps keep planning, tests, code review, commits, pull requests, and merging from becoming a manual checklist.

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/99xagency/godmodeskill/work
Clone the repo
git clone --depth 1 https://github.com/99xAgency/GodModeSkill
Per session 78 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 2,890 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 $0.00078 $0.02890
Opus 5 $0.00039 $0.01445
Sonnet 5 $0.00016 $0.00578
Haiku 4.5 $0.00008 $0.00289

Measured 2d ago against content hash f0aa16a7c0e2, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

work 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 2d 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.

skill/commands/work.md · 305 lines

How it starts

The opening of the file, as written. The whole thing — 305 lines — stays where its author put it; the contents beside it link to each section on GitHub.

/work — multi-LLM workflow

Single command for the four daily workflow shapes. Modes:

  • plan — propose, multi-LLM review, converge, write planning/<feature>.md
  • implement — design review → code → test → code review → PR → merge gate
  • major-bug — failing test (TDD) → fix-plan review → code → test → code review → PR
  • minor-bug — fast path: fix → test → commit (NO LLM gates)

The bash binary at ~/.local/bin/work is the executor. THIS command is the conductor — Claude orchestrates phases, asks the user where required, and never blindly merges.

Disambiguation — always ask if unclear

If the user request doesn't make the mode obvious, ask in numbered form:

Which workflow?
1. Plan a new feature (with multi-LLM review)
2. Implement (PR + multi-LLM code review + merge)
3. Major bug fix (failing test + multi-LLM review + PR)
4. Minor bug (just fix it, no review)

Hard rules (non-negotiable)

  1. Never hardcode agent names. Always call work pick-agents (LRU rotation across cdx-1/2/3, automatically uses cdx-4 / gem-2 if added later to agents.json).
  2. Lineage diversity required. Quorum needs ≥1 codex + ≥1 gemini + ≥1 opencode agreeing. Enforced by work-converge — Claude does NOT override the quorum rule lightly.
  3. /clear between different PRs. Handled automatically by work based on task_id change. Don't pass --clear flags by hand.
  4. Event-driven wait. work review blocks on inotifywait — zero polling, zero tokens during wait. Do NOT add monitoring loops in Claude.
  5. Auto-modify git for branch / commit / PR open / push. NEVER auto-merge — always ask user via the pre-merge checklist.
  6. All user clarification = numbered options, simple English.
  7. Disagreement loop: Claude (the proposer) revises, then re-runs work review --round N+1. Reviewers see prior rounds via --prior-rounds-file. Max 3 rounds before escalating to user.

Mode 1: plan

Use when user wants to design a new feature.

Read the full file on GitHub · 305 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. 2d ago First seen · 305 lines · 78 tokens per session scan A f0aa16a7c0e2

Subscribe to this mod's changes

work is a command published in the GitHub repository 99xAgency/GodModeSkill (241 stars, last pushed 3mo ago), licensed MIT. It adds 78 tokens to every session and 2,890 once invoked, about $0.0004 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.