suboss87/FDEOps

What it solves

FDEOps provides a persistent, local memory layer for AI coding agents. While AI agents often forget client-specific context, project goals, and stakeholder decisions between sessions, FDEOps creates a "fieldbook" (a private, local directory of Markdown files) that acts as a system of record for the entire engagement lifecycle.

How it works

FDEOps integrates as a skill or plugin for various AI coding agents (such as Claude Code, Cursor, or Codex). It uses a local CLI to manage a dedicated engagement folder (~/fde-engagements/<client>/) containing structured Markdown files for context, briefs, stakeholders, and decisions.

When a user interacts with the @fde command, the agent routes the conversation to update or retrieve information from these files. The system follows a specific field methodology (land $\rightarrow$ discover $\rightarrow$ plan $\rightarrow$ build $\rightarrow$ ship $\rightarrow$ close) to ensure that engagement details are captured as a side effect of the work.

Who it’s for

It is designed for Forward Deployed Engineers (FDEs), consultants, agency developers, solutions architects, and fractional CTOs who work with multiple clients and need a defensible, dated record of their engagements.

Highlights

• Agent Agnostic: The fieldbook lives on disk, allowing users to switch between different AI coding agents without losing client memory.
• Local-First Privacy: All data is stored in plain Markdown files locally; there are no network calls, telemetry, or accounts.
• Structured Methodology: Includes built-in phase verbs (land, discover, plan, etc.) to guide the engagement process.
• Fieldbook UI: Provides a local browser-based dashboard to view the current engagement status and records.
• Human-in-the-loop: The AI drafts updates to the record, but the user must confirm them before they are written to disk.