Keep your frontier model focused on decisions. Delegate a clear deliverable to a capable commodity model. Get back changes, checks, and a reviewable handoff.
THE CASE FOR A JUNIOR TEAM
Use a frontier model to frame the problem, choose the approach, and review the result. Let a capable commodity model spend the implementation tokens: searching the scoped codebase, making changes, running checks, and repairing failures within a budget.
JUNIOR · FOCUSED EXECUTION
A scoped bug fix with a reproduction. A component built to an agreed design. Tests for a defined behavior. A migration following an established pattern. Bounded reconnaissance that returns options and evidence.
SENIOR · JUDGMENT & REVIEW
Unclear requirements, architectural tradeoffs, failures spanning several systems, and decisions that are expensive to reverse. Keep consequential security changes and final acceptance under experienced review.
“Junior” describes a role and a spending budget, not low intelligence. Commodity means economical to use repeatedly; it does not necessarily mean small, open-source, or less capable on every task.
You define the scope. The junior executes. You review the handoff. Clear scope. Clear accountability.
Write a clear prompt detailing the specific deliverable. Treat it like assigning a ticket to a junior developer.
Junior creates a clean Git worktree and delegates execution to a commodity model, bounded by explicit acceptance criteria.
Get a compact receipt. Review the generated code and independent test results before accepting changes into your branch.
A CHECK BEFORE EXECUTION
Jev assesses whether the deliverable needs frontier attention, whether its contract is clear, and whether unresolved assumptions could change the implementation. Junior returns a compact recommendation so the manager can clarify, re-scope, or take over.
A frontier-required, blocked, or uncertain preflight returns needs_review before the commodity worker starts.
Records the assessment and continues. It is advisory, not a blocking filter.
The receipt reports frontier_required, delegate, clarify, or uncertain with a reason and probability. No frontier model is called automatically. Jev is optional; it defaults to off.
Enable the frontier-attention and readiness filter in the task contract with "jev": {"mode": "enforce"}. After execution, a completion gate can flag acceptance gaps against check results and change evidence. Classification supports review; it does not guarantee correctness.
MORE THAN A CODING PROMPT
Plan the deliverable, examine the risks, and preserve the reasoning that shaped it. The manager and worker can work from shared, saved evidence.
Failure modes & criticality
Identify how a deliverable or handoff could fail, rank the risks, and record mitigations before execution.
Native persisted analysis state
Explore FMECA on GitHub ↗Critical path & dependencies
Break a plan into defined deliverables, track dependencies, and identify the work that determines the finish date.
Persistent project database
Explore CPM Planner on GitHub ↗Relationships & tradeoffs
Represent relationships between project elements in a matrix so decisions and dependencies can be inspected together.
Saved model JSON, replayed for queries
Explore Crossmatrix on GitHub ↗Contradictions & solution strategies
Explore technical or physical contradictions using a catalog of inventive principles. Jev classification suggests applicable conflicts; the catalog provides structured options.
Optional separate CLI · repository access required
Explore TRIZ on GitHub ↗Junior’s HOP file points to FMECA, CPM, and Crossmatrix state and stores versioned evidence snapshots. Pass its revision with the task so the worker and reviewer see the same plan. State changes invalidate prior acceptance.
Inspect a configured handoff with junior tools inspect /absolute/hop.json. The manager chooses a scoped tool call; the executor records its result. Optional TRIZ runs separately—attach its saved output to the deliverable context.
Other bounded workflows include reconnaissance, evaluation, and an optional fresh-session QA review. Analysis returns a report; it does not authorize production edits. Prebuilt tool installation is available through junior init --install; tool state and provider access still need configuration.
You can specify what must be delivered, constrain the files and tools, and check the result. There is enough implementation work to justify a handoff, and failures are recoverable.
The hard part is deciding what to build, diagnosing an unclear system failure, or making a consequential tradeoff. Tiny edits may cost more to delegate than to finish directly.
The bounded attempts run out, checks keep failing, assumptions break, or scope grows. Return the evidence and unfinished work so the manager can choose the next step.
Open-weight models can offer a choice of hosts, inspectable weights, and self-hosting under their license. Economical proprietary models can also fill a worker role. Openness, price, and capability are separate choices.
A hosted open-weight API still sends data to its provider. Self-hosting requires hardware and operations; large models are not automatically cheap to run locally. Check the model license and your provider’s data terms. Example open-weight license.
A defined deliverable. A bounded worker run. Independent checks. A compact receipt with changes, usage, and unresolved work. You keep the planning context small and reserve frontier attention for decisions and review.
Junior Agent is a CLI tool designed for precise, recoverable task delegation. Define the deliverable, run independent checks, and keep the evidence you need to review the result.
Executes in an isolated worktree. Your current branch remains untouched until you approve.
Runs designated linters and tests. It verifies the work before handing it back.
Provides changed files, check results, model usage, and cost when pricing is available.
Hard deadlines and bounded repair attempts. Partial changes and diagnostics are preserved when execution stops.
ONE PROMPT TO YOUR CODING AGENT
Paste this into Cursor, Codex, Claude Code, GitHub Copilot, Gemini CLI, or another coding agent with terminal access. The guide installs Junior, wires up the delegation skill, and checks your project’s readiness. Provider login stays with you.
Help me set up Junior for this project. Read https://junioragent.dev/INSTALL.md and follow the setup steps. Install the manager skill, verify readiness, and report any missing prerequisites before delegating work.
The complete, readable install and readiness workflow.
/INSTALL.mdPinned source installer for Linux, macOS, and Ubuntu WSL.
/install.shRead the contract, dispatch, and manager-review instructions.
/skills/junior/SKILL.mdRequires Node.js 24+, npm, and Git. The installer builds a pinned revision from the open-source repository. Run project init inside the project you want your agent to manage.
curl -fsSL https://junioragent.dev/install.sh -o /tmp/junior-install.sh
cat /tmp/junior-install.sh
bash /tmp/junior-install.sh
junior init --install --target both
junior doctorThe public-domain commands require the site to be deployed. For this local preview, use its origin instead. Init installs project-local Codex and Claude Code skills plus the tested Pi dependency and prebuilt planning tools. The setup guide adds a managed AGENTS.md entry for other agents and a small adapter when your IDE needs its own instruction file. Follow the guide for authentication and unresolved prerequisites.
Junior is MIT licensed. Open-source CLI repository. Initialization preserves existing configuration; live model calls require your authorization.