Outsourcerer plugin
An orchestrator that splits work across your AI agents and models, keeps long jobs on disk and shows spend for each.
- 194 ★ on the repository
- Last commit
Broad access
Why this level
- Executors in other agents edit files and run commands on the lead session's order.
- The plugin spends subscription limits and paid OpenRouter keys.
- Context goes to cloud models on the chosen lanes.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add plugins/outsourcererDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewBroad accessThe tool can change data, run commands or handle secrets.
- Executors in other agents edit files and run commands on the lead session's order.
- The plugin spends subscription limits and paid OpenRouter keys.
- Context goes to cloud models on the chosen lanes.
This is third-party code. Review the repository files before installing.
What it does
Outsourcerer turns the current session into the lead: it hands routine work to a cheap model, hard calls to a strong one and private material to a model on your machine. Executors include Claude Code, Codex, Cursor, Droid, Antigravity, Devin, Hermes, Warp, Cline, Pi and Grok Bot. There is no server or framework; everything runs in plain speech. Big work lives on disk as a plan and survives a closed laptop. Advisors argue over the plan before building, a worker builds, another model reviews, and merges land one at a time only after your tests pass. `fanout` runs and loops retry on a cheap model until tests pass. A Tab shows real spend per lane. Skills can be passed to an executor whole.
Who it is for. People who already hold several subscriptions and keys for different models and want to send work to them from one session with a spend report.
Good fit when
- Repo mapping and test runs go to a cheap model while the main agent stays the lead.
- Several strong models review a complex plan and work starts only on consensus.
- A big build runs for days and after a restart the plan resumes where it stopped.
- A local Ollama model handles private tasks at zero dollars.
Not a fit when
- You have a single agent and no other models or subscriptions.
- You need a commercial licence for a company above the free threshold.
- The work cannot tolerate passing context to third-party cloud models.
Example request
Offload mapping the repository authentication to a cheap model and then show how much I saved against the main model.Limitations
The licence is PolyForm Noncommercial 1.0.0 with a written small-business permission: free for personal use, education and companies under five people with up to 20 thousand dollars of monthly revenue. Others need a paid licence. Keyless lanes spend your subscription limits, and paid OpenRouter models need a key. Models marked `:free` may train on prompts. Windows is supported through Git Bash with no WSL.
How to disable. Disable the plugin with `/plugin uninstall outsourcerer@outsourcerer` and for other agents remove links created by the `parity` commands. Delete plans and spend logs on disk by hand.
FAQ
How do I install it in Claude Code?
Run `/plugin marketplace add alexgreensh/outsourcerer`, then `/plugin install outsourcerer@outsourcerer`.
How do I connect Codex, Devin and Hermes?
For Codex `outsourcerer parity-codex`, for Devin `outsourcerer parity`, for Hermes `outsourcerer parity-hermes`. For Antigravity `agy plugin import claude-code` works.
Which driving modes exist?
Auto-pilot, you-drive and hybrid. The choice is remembered and changed with a plain phrase, and risk consent prompts hold in all three modes.
What stops key leaks?
Before cloud delegation a block fires if the working tree has a credential file such as `.env` or `id_rsa`. Work then runs in a clean copy without such files.
What does `suggest` do?
It shows which models are free or cheap right now on your platforms by reading catalogs live.
Discussion
No questions or reviews yet.
Related
PluginsA process kit for agents: a skill router, project memory in a DEV folder and result verification before saying done.
One agent launches others in the background: Claude Code, Codex, Copilot CLI, Grok Build and Pi, with a model chosen per stage.
A layer over coding agents that assembles a temporary specialist team per task inside Claude Code, Codex, Cursor and Gemini
A harness for long-running tasks: Claude Code or Codex becomes an orchestrator that launches workers and testers and decides when to stop
Hands routine work to Codex, Claude Code, Antigravity and Kimi workers via playbooks and reports what each did and cost.
Preloop is an open source agent control plane: an MCP firewall, a budgeted model gateway and event-driven agent runs.
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Repository stars | Main difference |
|---|---|---|---|
| Outsourcerer (this card) | no | 194 | |
| CLI Agent Orchestrator | no | 1,548 | an AWS orchestrator of several CLI agents; Outsourcerer routes work across models and platforms and keeps a spend ledger |
| Agency Orchestrator | no | 2,468 | agent role orchestration; here model choice per task, on-disk background plans and verification by your tests |
README in short
The plugin is scanned by repo-forensics before every release. Security notes cover a single key per lane read from `~/.env`, disclosure of the lane and tier on every dispatch and a prompt before destructive actions.
README badge
Original SKILL.md
---
name: outsourcerer
description: 'Delegate coding work to Codex, Cursor, Devin, Gemini, OpenRouter or local models with your skills. Use to outsource, offload, delegate, plan builds, get a second opinion, or run work in parallel.'
---
# Outsourcerer
**Created by**: [Alex Greenshpun](https://linkedin.com/in/alexgreensh)
Make your AI coding tools work as one team. Outsourcerer routes each job to the best-value capable lane available (plan-backed, keyless, free, local, or paid) while the current session stays in charge. It carries the user's skills, plugins, and MCP setup into delegated work, tracks limits and cash cost, and can bring in stronger models or multi-model review when the job calls for it. One entrypoint:
```
${CLAUDE_PLUGIN_ROOT}/skills/outsourcerer/scripts/outsourcerer.sh <subcommand> [args]
```
`--cloud-ack` and `--provider` work before or after `plan`/`task`/`land`/`flow`; a task's lane is stored from `task add ... --provider <lane>` or a leading `--provider`.
The user talks in plain language and you run the commands. They should never need to learn a flag, so run a command rather than explain it.
## You are the orchestrator
You are the captain of the squad, not a dispatcher waiting for orders. The user sets *direction*: where to go, what to do in general. Everything between that direction and the finished result is yours: the plan, the delegation, the verification, and the outcome.
Decide these yourself every time, without putting them to the user:
- Which lane, model, effort, and verb a task gets, and whether to delegate it at all.
- Branch, worktree, and version strategy. Which commit to cut from, what the next version number is.
- Task order, parallelism, and when to fan out.
- Cleanup of junk your own tooling produced: test litter, stale locks, stray job dirs, unrotated logs.
- Whether a result is good enough, or goes back to the delegate for another pass.
- Which bug to fix first when you find several.
Escalate only these three, and say which one it is:
- A fork in **direction** where the two paths are materially different work.
- Real cash spend, or cloud disclosure that has not been consented.
- Something irreversible that reaches beyond this machine (publishing, sending, force-pushing a shared branch, deleting the user's own data; your tool's leftovers don't count).
"I'm doing X because Y, say so if you disagree" is orchestration. "Should I do X, or Y?" hands the work back. When you catch yourself listing options for a call inside your remit, pick one and go. A concern about the ask gets one sentence, then you keep building.
A green checkmark from a delegate is a claim, not evidence. Verify before you report it, and report failures as loudly as wins.
## Session start: the handshake
Start every session by running `brief`. It returns the ready lanes, the user's live limits (e.g. `claude5h=95`), the exclude list, and a conservation call in one shot. Greet the user with that live picture. If their driving mode is unset, present the three modes and let them pick; the choice is remembered in `~/.outsourcerer/mode`, so ask once.