Lean Coder — Project Management agent for Claude Code
Token-efficient coding agent for any project.
How to install Lean Coder
Installs to ~/.claude/agents/rohithkh4b3kr3-multi-subagents-lean-coder.md
mkdir -p ~/.claude/agents && curl -fsSL https://raw.githubusercontent.com/rohithkh4b3kr3/multi-subagents/HEAD/agents/lean-coder.md -o ~/.claude/agents/rohithkh4b3kr3-multi-subagents-lean-coder.md Restart Claude Code, or start a new session, for it to be picked up.
What Lean Coder does
name: lean-coder model: sonnet description: Token-efficient coding agent for any project. Use for edits, refactors and bug fixes (multi-file or unfamiliar code) where reading raw files, logs or command output would burn context. Routes work through code-review-graph (structure), token-savior (symbols, memory), context-mode (big-output sandbox, indexed search) and rtk (compact shell output). Returns the same answer quality as a normal agent with far fewer tokens read.
You are a coding ag
Alternatives in Project Management
- Agent Evaluator — Evaluates agent output against 5-axis quality rubric (accuracy, completeness, clarity, actionability, concisen 243.5k ★
- Logs — Pod logs feature specialist 538 ★
- Pax — Senior Researcher 296 ★
Full documentation available on GitHub
View Source RepositoryRelated Agents
Task Initializer
Discovers project structure and initializes token-efficient task management system
Tp Onboard
PROACTIVELY use this when the user is exploring an unfamiliar codebase — asks "how is this organised", "what d
Tp Run
PROACTIVELY use this general-purpose token-pilot workhorse for ANY coding task (read / edit / search / explore
Codex Analyst
Token-heavy READ-ONLY analysis, codebase surveys, audits and inventories, executed by the Codex CLI rather tha
Conductor
Primary orchestrator. Routes every task to the matching specialist subagent. Cannot write or edit files direct
Biz Model
Business analysis and business model proposal. Use to define a business model, create a Lean Canvas, or evalua
Related Skills
Engineering Craft
Use when writing or changing any production code or tests — new features, bug fixes, refactors, or test author
Anvil
Anvil is a single, self-contained AI coding agent for the terminal. The coder is a real agent — it reads, writ
Instruction Health Skills
Agent skills to keep AI instruction files (AGENTS.md, CLAUDE.md, MEMORY.md) lean. Guardian runs before edits t