Oma Coordination
Step-by-step plan for splitting a multi-domain feature across PM, dev and QA agents
Test report
- Verdict
- Works with setup
- Score
- Tested
- Jul 21, 2026
- Environment
- Claude Code 2.x (agent harness)
- Upstream re-checked
- Aug 10, 2026 · 8272964
Cloned the repo and installed generated/agent-skills/oma-coordination into a throwaway HOME — frontmatter parsed to exactly {name, description}, and the one referenced file (resources/examples.md) fetched HTTP 200; no curl|sh, base64 or secret-handling anywhere in the skill body. I ran the real CLI sandboxed (`bunx oh-my-agent@latest agent:spawn --help` in a temp HOME) and confirmed the four flags the SKILL.md cites — -m/--model, -w/--workspace, --isolation worktree, --read-only — all exist as documented; I did not execute an actual spawn, since that needs an authenticated vendor CLI, and `oma` is not on PATH after a plain skill copy. Baseline vs skill on the same task (add real-time notifications to a React web + Flutter mobile + FastAPI app): baseline was 227 words of prose with 0 executable commands, no PM step, and "agree on the JSON shape early" as advice; the skill artifact was 780 words with 8 concrete `oma agent:spawn` invocations, a session id, a PM decomposition producing 6 tasks in 4 priority tiers, a hard P1→P2 gate that blocks web/mobile until the backend contract is frozen, `--isolation worktree` on the two parallel spawns, named progress-{agent}-{sessionId}.md files to poll, and a --read-only QA spawn last. Docked output and docs because roughly a third of the skill artifact is schema boilerplate (SSL primitives table, resource-scope table) that changed no decision, and because SKILL.md lists sibling agents and the CLI as "Dependencies" without saying that a bare skill copy leaves every emitted command as command-not-found.
Scored on four weighted criteria — install, triggering, output vs. baseline, docs. How scoring works
- Installs cleanly 5/5
- Triggers reliably 5/5
- Output vs. baseline 6/10
- Docs & honesty 4/5
What Oma Coordination does
Turns a multi-domain feature request into a manual multi-agent coordination sequence: PM decomposition first, tasks grouped into priority tiers, same-tier agents spawned in parallel into separate workspaces, API contract frozen before frontend/mobile work starts, and a QA pass last. Triggers when a task spans backend, frontend, mobile and QA and the user wants step-by-step control over agent spawn order rather than hands-off automation. Emits `oma agent:spawn` commands and names the progress files to poll; for fully automated execution it defers to the oma-orchestrator skill.
How to install Oma Coordination
git clone --depth 1 https://github.com/first-fluke/oh-my-agent.git /tmp/oma-coordination-src
mkdir -p ~/.claude/skills
cp -R /tmp/oma-coordination-src/generated/agent-skills/oma-coordination ~/.claude/skills/oma-coordination
# The skill body emits `oma agent:spawn ...` commands. That CLI is NOT installed by the copy above.
# To make the emitted commands runnable, install the harness in your project (needs bun + uv):
# cd /path/to/your/project && bunx oh-my-agent@latest
# Verified flags used by this skill: -m/--model <vendor>, -w/--workspace <path>,
# --isolation worktree, --read-only (confirmed via `oma agent:spawn --help`).
# The skill also assumes sibling agents exist: pm, backend, frontend, mobile, qa, orchestrator.
# Get them all with: cp -R /tmp/oma-coordination-src/generated/agent-skills/oma-* ~/.claude/skills/
# Alt distribution (skills only, no CLI): apm install first-fluke/oh-my-agent
Skills live in ~/.claude/skills/ (global) or .claude/skills/
(per-project). Restart Claude Code after installing.
Commands — how to trigger Oma Coordination
-
/oma-coordinationStep-by-step plan for splitting a multi-domain feature across PM, dev and QA agents
It also activates on plain-language prompts like these:
-
Coordinate the frontend and backend agents on this task -
Walk me through assigning work to QA agents -
Help me manage this multi-agent project workflow
Frequently asked questions
- Is the Oma Coordination skill free?
- Yes. The skill itself is free from first-fluke/oh-my-agent. SkillProof publishes the install command and an independent test verdict at no cost.
- Does Oma Coordination work with Claude Code?
- We tested it with Claude Code 2.x (agent harness) on Jul 21, 2026. Verdict: Works with setup. Cloned the repo and installed generated/agent-skills/oma-coordination into a throwaway HOME — frontmatter parsed to exactly {name, description}, and the one referenced file (resources/examples.md) fetched HTTP 200; no curl|sh, base64 or secret-handling anywhere in the skill body. I ran the real CLI sandboxed (`bunx oh-my-agent@latest agent:spawn --help` in a temp HOME) and confirmed the four flags the SKILL.md cites — -m/--model, -w/--workspace, --isolation worktree, --read-only — all exist as documented; I did not execute an actual spawn, since that needs an authenticated vendor CLI, and `oma` is not on PATH after a plain skill copy. Baseline vs skill on the same task (add real-time notifications to a React web + Flutter mobile + FastAPI app): baseline was 227 words of prose with 0 executable commands, no PM step, and "agree on the JSON shape early" as advice; the skill artifact was 780 words with 8 concrete `oma agent:spawn` invocations, a session id, a PM decomposition producing 6 tasks in 4 priority tiers, a hard P1→P2 gate that blocks web/mobile until the backend contract is frozen, `--isolation worktree` on the two parallel spawns, named progress-{agent}-{sessionId}.md files to poll, and a --read-only QA spawn last. Docked output and docs because roughly a third of the skill artifact is schema boilerplate (SSL primitives table, resource-scope table) that changed no decision, and because SKILL.md lists sibling agents and the CLI as "Dependencies" without saying that a bare skill copy leaves every emitted command as command-not-found.
- What is the Oma Coordination SkillProof Score?
- 8.0/10 — installs cleanly 5/5, triggers reliably 5/5, output vs. baseline 6/10, docs & honesty 4/5.
- How do I install Oma Coordination?
- Copy the install command from this page, run it in your terminal, and restart Claude Code. Skills live in ~/.claude/skills/ (global) or .claude/skills/ inside a project.
- Can I use Oma Coordination with Cursor, Copilot, Gemini CLI, Codex or other AI tools?
- The SKILL.md format is native to Claude (Claude Code, Desktop, claude.ai). The instructions inside adapt to other assistants: Cursor rules, GitHub Copilot instructions, Windsurf rules, Custom GPTs, AGENTS.md for OpenAI Codex, and GEMINI.md for Google Gemini CLI — our conversion guides cover each, and the free converter on the tools page does the wrapping for you.