haft

Name: haft
Author: m0n0x41d

1.4kstars103forks

How to install haft?

m0n0x41d/haft

$git clone https://github.com/m0n0x41d/haft

Installs into the current project.

Summary

*formerly quint-code*

Files · 1

View on GitHub

README.md

1	<img src="assets/banner.svg" alt="Haft" width="600">
2
3	formerly [quint-code](https://github.com/m0n0x41d/quint-code)
4
5	FPF governance substrate for AI-assisted software delivery.
6
7	Your agents (Claude Code, Codex) write code fast. Most repositories are not ready
8	for serious harness engineering: the target system is underspecified, the
9	enabling system is implicit, term maps are missing, and runtime evidence is
10	detached from the spec. Haft makes the project harnessable before it scales
11	execution.
12
13	---
14
15	## What is Haft?
16
17	Haft is a governance substrate that makes a repository harnessable for
18	principal-led FPF engineering work. It turns problem frames, comparisons,
19	decisions, commissions, and evidence into auditable artifacts, with enforcement
20	at the kernel boundary.
21
22	Specify → Think → Run → Govern.
23
24	Not a coding agent. Not a documentation generator. The handle between the tool
25	and the hand: the part that turns raw model capability into formal
26	specification, governed decisions, bounded commissions, and evidence-backed
27	engineering work.
28
29	### Three surfaces, one artifact graph
30
31	Haft is consumed through three surfaces over one `.haft/` artifact graph:
32
33	- Skills + slash commands in your coding agent (Claude Code, Codex, OpenCode, Cursor) — workflow skills auto-trigger; `/h-frame /h-decide /h-verify ...` run manually
34	- CLI (`haft problem`, `haft solution`, `haft decision`, ...) — manual access, no LLM in the loop
35	- MCP server (`haft serve`) — programmatic access for any LLM agent over the Model Context Protocol
36
37	The kernel MCP server is the cross-host enforcement surface: it validates
38	arguments server-side and returns structured errors for FPF violations (missing
39	required fields, parity gaps, weakest-link omissions, predictions without
40	verify_after). Skills carry the procedure; the kernel carries the gates.
41
42	### What changed in v8
43
44	v8 dropped the standalone interactive agent (`haft agent`), the TUI, and the
45	desktop wrappers. Haft no longer competes with general coding agents on the
46	runtime surface — it adds governance discipline on top of whichever agent you
47	already use. The pivot, with parity-compared variants, rollback plan, and
48	falsifiable predictions, is recorded in
49	`.haft/decisions/dec-20260525-v8-architecture-pivot-from-standalone-agent-to-g-bbe45cb7.md`.
50
51	Upgrading from v7? See [MIGRATION-v8.md](MIGRATION-v8.md) — the upgrade checklist
52	plus what was dropped (`haft agent`, TUI, desktop, v7 helper commands).
53
54	---
55
56	## Built on First Principles Framework
57
58	[FPF](https://github.com/ailev/FPF) by
59	[Anatoly Levenchuk](https://www.linkedin.com/in/ailev/) — a rigorous,
60	transdisciplinary architecture for thinking.
61
62	The skill set (`h-frame`, `h-explore`, `h-compare`, `h-decide`, `h-verify`, and
63	the full catalog below) gives your agent an FPF-native operating system for
64	engineering decisions: framing before solutions, characterization before
65	comparison, parity enforcement, evidence with congruence penalties,
66	weakest-link assurance, and a cycle that reopens itself when evidence ages or a
67	measurement fails.
68
69	The framing and comparison skills auto-trigger on operator context. The binding
70	step (`h-decide`, `h-commission`) is manual-only per the Transformer Mandate:
71	agents frame and compare; the human principal records the binding choice.
72
73	`haft fpf search` (and `haft_query(action="fpf")` from MCP) searches the indexed
74	FPF specification. Retrieval is hybrid: exact pattern id first, then keyword
75	(FTS5) fused with semantic recall over baked section vectors, so a reworded "how
76	do I think about X" finds the pattern that answers it. The vectors ship inside
77	the binary; semantic recall degrades to keyword when the embedding sidecar is
78	absent.
79
80	---
81
82	## Install
83
84	```bash
85	curl -fsSL https://raw.githubusercontent.com/m0n0x41d/haft/main/install.sh \| bash
86	```
87
88	The install URL still points at the historical `quint-code` path. The installed
89	binary is `haft`.
90
91	Then in your project, init with your host-agent flag:
92
93	```bash
94	haft init

People who liked this also viewed

6 picks

Type

clickup-mcp-serverClickUp MCP Server - Integrate ClickUp project management with AI through Model Context ProtocolMCP ServersJul 202694k49
plane-mcp-serverPlane's Official Model Context Protocol Server 🔌 ⌨️ 🔥MCP ServersJul 202621k269
linear-mcp-serverA server that integrates Linear's project management system with the Model Context Protocol (MCP) to allow LLMs to interact with Linear.MCP ServersMay 20253.8k346
intervals-icu-mcpRead/write MCP server for Intervals.icu — training data, planning, and structured workout generation.MCP ServersJul 20263.7k34
yandex-tracker-mcpYandex Tracker MCP Server with OAuth2 supportMCP ServersJul 20262.5k91
ham.runOpen-source marathon physiology calculators for endurance runners. 12 tools: race time prediction (Riegel model), pacing strategy, fueling plans, running economy analysis, caffeine protocols, heat acclimation planning, periodization compariMCP Serversrecently2.5k0

haft

1.4kstars103forks

View on GitHub

How to install haft?

m0n0x41d/haft

$git clone https://github.com/m0n0x41d/haft

Installs into the current project.

Summary

*formerly quint-code*

Files · 1

View on GitHub

README.md

1	<img src="assets/banner.svg" alt="Haft" width="600">
2
3	formerly [quint-code](https://github.com/m0n0x41d/quint-code)
4
5	FPF governance substrate for AI-assisted software delivery.
6
7	Your agents (Claude Code, Codex) write code fast. Most repositories are not ready
8	for serious harness engineering: the target system is underspecified, the
9	enabling system is implicit, term maps are missing, and runtime evidence is
10	detached from the spec. Haft makes the project harnessable before it scales
11	execution.
12
13	---
14
15	## What is Haft?
16
17	Haft is a governance substrate that makes a repository harnessable for
18	principal-led FPF engineering work. It turns problem frames, comparisons,
19	decisions, commissions, and evidence into auditable artifacts, with enforcement
20	at the kernel boundary.
21
22	Specify → Think → Run → Govern.
23
24	Not a coding agent. Not a documentation generator. The handle between the tool
25	and the hand: the part that turns raw model capability into formal
26	specification, governed decisions, bounded commissions, and evidence-backed
27	engineering work.
28
29	### Three surfaces, one artifact graph
30
31	Haft is consumed through three surfaces over one `.haft/` artifact graph:
32
33	- Skills + slash commands in your coding agent (Claude Code, Codex, OpenCode, Cursor) — workflow skills auto-trigger; `/h-frame /h-decide /h-verify ...` run manually
34	- CLI (`haft problem`, `haft solution`, `haft decision`, ...) — manual access, no LLM in the loop
35	- MCP server (`haft serve`) — programmatic access for any LLM agent over the Model Context Protocol
36
37	The kernel MCP server is the cross-host enforcement surface: it validates
38	arguments server-side and returns structured errors for FPF violations (missing
39	required fields, parity gaps, weakest-link omissions, predictions without
40	verify_after). Skills carry the procedure; the kernel carries the gates.
41
42	### What changed in v8
43
44	v8 dropped the standalone interactive agent (`haft agent`), the TUI, and the
45	desktop wrappers. Haft no longer competes with general coding agents on the
46	runtime surface — it adds governance discipline on top of whichever agent you
47	already use. The pivot, with parity-compared variants, rollback plan, and
48	falsifiable predictions, is recorded in
49	`.haft/decisions/dec-20260525-v8-architecture-pivot-from-standalone-agent-to-g-bbe45cb7.md`.
50
51	Upgrading from v7? See [MIGRATION-v8.md](MIGRATION-v8.md) — the upgrade checklist
52	plus what was dropped (`haft agent`, TUI, desktop, v7 helper commands).
53
54	---
55
56	## Built on First Principles Framework
57
58	[FPF](https://github.com/ailev/FPF) by
59	[Anatoly Levenchuk](https://www.linkedin.com/in/ailev/) — a rigorous,
60	transdisciplinary architecture for thinking.
61
62	The skill set (`h-frame`, `h-explore`, `h-compare`, `h-decide`, `h-verify`, and
63	the full catalog below) gives your agent an FPF-native operating system for
64	engineering decisions: framing before solutions, characterization before
65	comparison, parity enforcement, evidence with congruence penalties,
66	weakest-link assurance, and a cycle that reopens itself when evidence ages or a
67	measurement fails.
68
69	The framing and comparison skills auto-trigger on operator context. The binding
70	step (`h-decide`, `h-commission`) is manual-only per the Transformer Mandate:
71	agents frame and compare; the human principal records the binding choice.
72
73	`haft fpf search` (and `haft_query(action="fpf")` from MCP) searches the indexed
74	FPF specification. Retrieval is hybrid: exact pattern id first, then keyword
75	(FTS5) fused with semantic recall over baked section vectors, so a reworded "how
76	do I think about X" finds the pattern that answers it. The vectors ship inside
77	the binary; semantic recall degrades to keyword when the embedding sidecar is
78	absent.
79
80	---
81
82	## Install
83
84	```bash
85	curl -fsSL https://raw.githubusercontent.com/m0n0x41d/haft/main/install.sh \| bash
86	```
87
88	The install URL still points at the historical `quint-code` path. The installed
89	binary is `haft`.
90
91	Then in your project, init with your host-agent flag:
92
93	```bash
94	haft init

People who liked this also viewed

6 picks

Type