Feos Builder — Project Management agent for Claude Code
FEOS S5 implementation agent.
How to install Feos Builder
Installs to ~/.claude/agents/mjtpena-frontier-engineering-os-feos-builder.md
mkdir -p ~/.claude/agents && curl -fsSL https://raw.githubusercontent.com/mjtpena/frontier-engineering-os/HEAD/.claude/agents/feos-builder.md -o ~/.claude/agents/mjtpena-frontier-engineering-os-feos-builder.md Restart Claude Code, or start a new session, for it to be picked up.
What Feos Builder does
name: feos-builder description: "FEOS S5 implementation agent. Use to implement exactly one backlog task (TK-###) on its own branch, making the Tester's failing tests pass without modifying them, following all instruction layers." tools: Read, Grep, Glob, Write, Edit, Bash model: sonnet
FEOS Builder (stage S5)
**Mission:** Implement one task, minimally and correctly, so that the specification-derived tests pass and every standard is met.
Scope
- Reads:
plan/backlog.yaml, `a
Alternatives in Project Management
- Keymappings — Keymappings specialist 538 ★
- Iwsdk Code Reviewer — Reviews IWSDK code for performance, ECS patterns, and project conventions 349 ★
- Agent Protocol & Architectural Mandate — Version: 2.10.5 Target Project: git-mcp-server Last Updated: 2026-03-25 This document defines the operational 206 ★
Full documentation available on GitHub
View Source RepositoryRelated Agents
Feos Planner
FEOS S4 Planning. Use to decompose an approved PRD and architecture into a schema-valid backlog (epics, featur
Release Manager Agent
Verifies the implementation against every acceptance criterion in the task file. Updates CHANGELOG.md on a ful
Fullstack Dev Agent
Implements features exactly as specified in the task file — no more, no less. Creates the feature branch, sati
Worktree Implementer
Default worker for one lane of a parallel wave. Use when dispatching a single, well-scoped implementation or f
Nextjs Implementer
Implements exactly one approved Next.js SDD task with tests. Use only after specification and planning gates p
Grainulator
Autonomous research sprint subagent. Executes multi-pass research, adds claims, compiles, and generates output
Related Skills
TDD Green
Implement minimal code to make failing tests pass in TDD green phase:
Theoretical Economics Claude Skill
Read the file SKILL.md at the root of this project and follow every instruction in it exactly.
Context Explain
Show exactly which instruction files each coding agent loads for a file or directory.