# 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.

Repository: https://github.com/m0n0x41d/haft
Canonical: https://ross.abutalabs.com/products/haft
Homepage: https://haft.tools
Language: Go
License: NOASSERTION
License Family: other
Topics: ai-agents, ai-coding, anthropic, claude, claude-code, developer-tools, first-principles, reasoning, cursor, cursor-ai, gemini-cli, context-memory, codex, human-ai-collaboration, mcp, systems-engineering, spec-driven-development
Last push: 2026-08-11T16:54:17+00:00

## Health v2 (maintenance only)
Score: 81/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 97, release rhythm 97, longevity 19
- inputs: {"age_days": 266, "days_push": 22, "days_rel": 22, "gap_med": 2, "n_releases_24m": 28}
- flags: no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1387, forks 102 (observed 2026-08-28T04:04:35.122748+00:00)

## What it is
Haft is a local governance and project-memory layer for AI coding agents such as Claude Code, Cursor, Codex, and Gemini CLI. Built on the First Principles Framework (FPF), it records framed problems, compared options, human decisions, and supporting evidence in a typed project graph, surfacing stale evidence and specification drift over time.

## Use cases
- preserve engineering decision rationale across AI coding sessions
- make my AI agent compare alternatives before picking a solution
- detect when past decisions are based on stale evidence
- add MCP governance gates to Claude Code or Codex
- keep a durable project memory of specs and decisions
- migrate AI-assisted project decisions out of chat history
- enforce parity when comparing design options with an AI agent

## When to choose
- your project will outlive the current chat session and decisions must be auditable
- multiple people or sessions need to rely on recorded decisions and evidence
- you want structured reasoning (framing, comparison, evidence decay) inside AI coding tools
- you use Claude Code, Codex, Cursor, or Gemini CLI and want MCP-based project memory

## When to avoid
- you only need lightweight TODO or task tracking without decision governance
- your work is short-lived scripts or throwaway prototypes where rationale loss is cheap
- you want a hosted or team-synced solution - Haft is local, single-project storage
- you need built-in task execution - v9 removed haft run and the bundled runtime

## Facets
- artifact type: cli-tool
- maturity: active
- function: mcp, agent-framework, developer-tools, cli, database, documentation
- domain: developer-tools
- platform: cli, cross-platform, windows, go
- tags: ai-coding-agents, decision-engineering, project-memory, first-principles-framework, spec-driven-development, evidence-decay, governance, claude-code, cursor, codex, gemini-cli, sqlite, ai-agents, automation, command-line, macos, linux

## Member repositories
- m0n0x41d/haft (main) score 81

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:35.122748+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T04:39:50.361493+00:00, confidence not recorded.
  - readme: https://github.com/m0n0x41d/haft (fetched 2026-08-28T04:04:35.122748+00:00, sha 246f2164f151)
  - homepage: https://haft.tools (fetched 2026-08-29T11:55:15.581347+00:00, sha ed0848476c80)
  - site_page: https://haft.tools/docs/migration (fetched 2026-08-29T11:55:15.592518+00:00, sha db6792b89d18)
  - site_page: https://haft.tools/docs (fetched 2026-08-29T11:55:15.590706+00:00, sha a8108ce96baa)
- Data as of 2026-08-30T08:39:29.467469+00:00.