Your AI Coder Needs a Brief. This Is It.

What it does

FDEOps is a Node.js CLI and skill pack for AI coding agents that structures the operational work surrounding generated code. It creates a per-client folder of markdown records—brief, stakeholders, decisions, risks, delivery evidence—and exposes 30 situational skills (from /brief to /close) that guide the agent through scoping, sign-offs, staging proofs, and handover. A local dashboard gives a read-only portfolio view of what needs attention across active clients.

The interesting bit

The project treats the AI agent as a junior consultant who needs structured supervision. Instead of asking the agent to write code faster, FDEOps gives it project-hygiene skills—like interrogating the brief, rehearsing rollback, or preparing a Friday readout—while keeping the canonical engagement record in plain markdown on your laptop, separate from the client repo.

Key highlights

• 30 situational skills mapped to six stages: Land, Discover, Plan, Ship, Outcome, Close
• Per-client .fde/ folders in plain markdown: brief.md, decisions.md, risks.md, delivery.md, and others
• Read-only HTML dashboard showing next actions, missing evidence, and delivery gaps across engagements
• Claude Code plugin, Cursor adapter, and generic skill support for Codex and other agents
• Local-first by design: records stay on your laptop, not in the client repository or a cloud service

Caveats

• The debrief --smart phrase recognition can miss details that need manual correction
• The default 16 KiB summary cap limits FDEOps output, not necessarily all agent context
• Several referenced features (MCP recipes, industry overlays) are mentioned but not deeply detailed in the README

Verdict

Grab it if you are a forward-deployed engineer or consultant using AI agents and need to track who signed off on what. If your work has no client handoffs or operational transfer, the authors themselves suggest a simpler note will do.