m0n0x41d/haft

Engineering decisions engine that know when they're stale. Frame, compare, decide — with evidence decay and parity enforcement. For Claude Code, Cursor, Gemini CLI, Codex and more.

What it solves

Haft provides a local governance layer for AI coding agents, preventing the loss of critical project context. It ensures that the "memory" of a project—such as the problem being solved, compared options, human decisions, and supporting evidence—is durable and structured rather than being lost in transient chat histories.

How it works

Haft acts as an intermediary between the human, the AI agent, and the repository. It implements the First Principles Framework (FPF), a rigorous system for engineering thinking, and provides this framework as a versioned source for the agent to query.

It operates through three primary surfaces:

  • MCP Server: A Model Context Protocol server (haft serve) that provides programmatic access and enforces "kernel gates" (server-side validation of required fields and authority boundaries).
  • Skills: Specialized procedures integrated into agents (e.g., /h-frame for clarifying problems, /h-decide for binding decisions) that guide the agent's workflow.
  • CLI: Direct manual access to the project's artifact graph and memory via commands like haft artifact and haft spec.

Project data is stored locally in a SQLite database and markdown carriers in a .haft/ directory.

Who it’s for

Software engineers using AI coding agents (such as Claude Code, Codex, Grok, or Cursor) who need a formal, durable way to track engineering decisions, specifications, and project governance within their codebase.

Highlights

  • Reliance-bearing memory: Distinguishes between casual chat reasoning and formal project records that future work must rely on.
  • Kernel gates: Enforces structural and authority boundaries server-side rather than relying on prompt-based discipline.
  • Evidence decay: Tracks when evidence for a decision has gone stale and surfaces it for review.
  • Source-native FPF delivery: Allows agents to retrieve versioned FPF source material with provenance.
  • Human-in-the-loop: Ensures binding decisions and commissions remain the explicit authority of the human principal.

Related

  • Project
    ailev/FPF

    A standards-style pattern language for turning complex engineering and research work into explicit, reviewable reasoning to maintain coherence across teams and AI agents.

  • Project
    AVIDS2/memorix

    Memorix is a Node‑JS CLI/library that adds a local, SQLite‑backed, searchable memory store to any Git project. It lets AI coding assistants (Claude Code, Copilot, Cursor, etc.) share project facts, code snapshots, design rationales and commit‑derived knowledge across sessions and tools via the Model‑Context‑Protocol. Features include autopilot worksets, observation/reasoning memory, Git‑memory ingestion, code‑graph indexing, orchestration of multi‑agent tasks, a web dashboard, and optional sync. Install with npm, run `memorix setup --agent <agent>` and the agents can start reusing the shared memory.

  • Project
    riponcm/projectmem

    projectmem adds a local, file‑based memory layer for AI coding assistants. It records typed events (issues, attempts, fixes, decisions) in a hidden `.projectmem/` folder, serves them via a single MCP server, and provides git‑hook pre‑commit warnings and visual dashboards so agents can reuse past experience and avoid repeating failures.

  • Project
    mcncarl/agent-memory-vault

    Agent Memory Vault is a Python toolkit that provides a local, Git‑backed Markdown vault with SQLite (and optional vector) search, enabling Claude Code, Codex and similar AI agents to share durable, auditable long‑term memory across sessions.

  • Project
    agnt-gg/agnt

    AGNT is a local‑first, self‑hostable desktop app that bundles AI agents, visual workflows, long‑running goals, memory, plugins and a local HTTP API into one “agent operating system”. Built with Electron, Vue, Express and SQLite, it supports 15+ LLM providers, 60+ workflow nodes, plugin templates and SkillForge for self‑improving skills. Ideal for developers or teams that need durable, inspectable AI automation on‑premise.