gm-skill 2.0.2232 → 2.0.2233

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
package/README.md CHANGED
@@ -1,27 +1,48 @@
1
1
  # glootius maximus (gm)
2
2
 
3
- **the coding agent skill that refuses to stop early or ship a stub.**
3
+ **Your coding agent doesn't decide when it's done. A gate does.**
4
4
 
5
- 14000+ hours of supervised modification, 8600+ commits, one person, built to raise one number: the signal-to-noise ratio of a coding agent. gm convinces your agent it already is a deterministic state machine (`PLAN -> EXECUTE -> EMIT -> VERIFY -> CONSOLIDATE -> COMPLETE`) and enforces that conviction with a wasm-backed orchestrator, witnessed execution, and a covering family of bounded subsets that refuses to let "follow-up" become a synonym for "I gave up."
5
+ ```
6
+ $ transition to=COMPLETE
7
+
8
+ DENIED CONSOLIDATE -> COMPLETE 2 residuals
9
+
10
+ x worktree-clean 3 uncommitted files
11
+ x ci-validated-fresh .ci-validated sha 7c90878 != HEAD e8ea29f
12
+
13
+ next: git_finalize
14
+ ```
15
+
16
+ Most agent harnesses ask the model to follow a process. gm makes the process a program the model cannot talk its way past.
6
17
 
7
18
  [![npm](https://img.shields.io/npm/v/gm-skill.svg)](https://www.npmjs.com/package/gm-skill) [![license](https://img.shields.io/badge/license-MIT-blue.svg)](./LICENSE) [![Discord](https://img.shields.io/badge/discord-join-5865F2.svg)](https://discord.com/invite/c9VV59MKNr) [![site](https://img.shields.io/badge/site-anentrypoint.github.io%2Fgm-informational.svg)](https://anentrypoint.github.io/gm/)
8
19
 
9
- **[Install](#install)** · **[How it works](#how-it-works)** · **[Release pipeline](#release-pipeline)** · **[Developing gm itself](#developing-gm-itself)** · **[Full paper (site)](https://anentrypoint.github.io/gm/paper/)** · **[Discord](https://discord.com/invite/c9VV59MKNr)** · **[License](#license)**
20
+ **[Why gm hits different](#why-gm-hits-different)** · **[Install](#install)** · **[How it works](#how-it-works)** · **[Release pipeline](#release-pipeline)** · **[Developing gm itself](#developing-gm-itself)** · **[Full paper (site)](https://anentrypoint.github.io/gm/paper/)** · **[Discord](https://discord.com/invite/c9VV59MKNr)** · **[License](#license)**
21
+
22
+ ```
23
+ npx gm-skill install
24
+ ```
25
+
26
+ ## Why gm hits different
27
+
28
+ **The COMPLETE gate is code, not a prompt.** Seven conditions guard the last transition -- PRD closed, mutables resolved, worktree clean, residual scan fired, CI green against the current HEAD sha, browser witness coverage, submodules in sync. They are a `Vec<String>` on an edge in `fsm.rs`, evaluated in Rust. A failed gate refuses the transition. The agent cannot narrate its way to done.
10
29
 
11
- it is named after **glootius maximus**, the muscle that holds you in the chair while you finish the work. the name is the joke and the discipline at once: the agent that sits down through PLAN -> EXECUTE -> EMIT -> VERIFY -> CONSOLIDATE -> COMPLETE actually ships. the agent that stands up early ships a stub with a green check on it.
30
+ **A refused transition tells the agent what to do next.** Gate denials carry the recovery verb: `worktree-clean` returns `git_finalize`, `ci-validated-fresh` returns `exec_js`. You get an instruction, not an error.
12
31
 
13
- that orientation is also why gm is built for token austerity: every token an agent spends should be signal toward the work, never narration, hedging, or busy-output. austerity is SNR enforced at the budget.
32
+ **Repeat the same failure and it stops you.** After the same denial fires repeatedly, the response stops restating the refusal and instructs the agent to record the stuck state and switch to a bounded-retry discipline. Loops end.
14
33
 
15
- free, open source, maintained by one person.
34
+ **Zero test files, and that is checkable.** Search this repo for `*.test.*`, `*.spec.*`, `__tests__`, or a jest config. There are none, and there never will be. Verification means running the real path and reading the real output in the same turn. VERIFY also greps the diff for `Mock*`/`Fake*`/`Stub*` -- a mock shipped as a real integration is the same violation as a test file.
16
35
 
17
- disclaimer: this is extremely opinionated. it will block bash, redirect your tools, refuse to write test files, force you to push git before ending a session, and reject any execute call without an explicit timeout. if that sounds terrible, this is not for you. if that sounds like what you wish your agent did automatically, keep sitting down.
36
+ **You can loosen the rules, and it will tell on you.** The phase graph is JSON at `.gm/instructions/fsm/graph.json`. Rewire edges, add states, swap gates. gm compares your graph against the compiled default and reports every edge you made weaker.
37
+
38
+ This is extremely opinionated. It narrows bash to a handful of prefixes, routes git through verbs, refuses to write test files, forces a push before a session ends, and rejects any execute call without an explicit timeout. If that sounds terrible, this is not for you. If that sounds like what you wish your agent did automatically, keep sitting down.
39
+
40
+ 14000+ hours of supervised modification, 8800+ commits, one person. Free, open source. Named after **glootius maximus**, the muscle that holds you in the chair while you finish the work.
18
41
 
19
42
  ## install
20
43
 
21
44
  A Claude Code Agent Skill is just a directory at `~/.claude/skills/<name>/SKILL.md` (personal, all projects) or `.claude/skills/<name>/SKILL.md` (one project). The directory name becomes the slash command. No marketplace, no `npx skills` library -- the installer copies the directory into place.
22
45
 
23
- **The npm package is `gm-skill`, not `gm`.** `npx gm@latest` resolves to an unrelated, unmaintained GraphicsMagick wrapper (`gm` on npm, last published years ago) with no `install` command -- it fails with `could not determine executable to run`. Always spell out the full package name below.
24
-
25
46
  Interactive (offers Claude Code settings):
26
47
 
27
48
  ```
@@ -40,6 +61,8 @@ Project-local instead of home (`./.claude/skills/gm`):
40
61
  npx gm-skill install --project
41
62
  ```
42
63
 
64
+ Spell out the full package name: the npm package is `gm-skill`, not `gm`. `npx gm@latest` resolves to an unrelated, unmaintained GraphicsMagick wrapper with no `install` command and fails with `could not determine executable to run`.
65
+
43
66
  The skill installs as `/gm`. On Claude Code the installer also offers (interactive) or sets (`--yes`):
44
67
 
45
68
  - `autoCompactEnabled: true`
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "gm-plugkit",
3
- "version": "2.0.2232",
3
+ "version": "2.0.2233",
4
4
  "description": "Bootstrap and daemon-spawn tool for gm plugkit binary. Downloads the correct platform wasm, verifies SHA256, and launches agentplug-runner (the native wasm host) as the spool watcher daemon.",
5
5
  "main": "index.js",
6
6
  "bin": {
package/gm.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "gm",
3
- "version": "2.0.2232",
3
+ "version": "2.0.2233",
4
4
  "description": "Spool-dispatch orchestration engine with unified state machine, skills, and automated git enforcement",
5
5
  "author": "AnEntrypoint",
6
6
  "license": "MIT",
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "gm-skill",
3
- "version": "2.0.2232",
3
+ "version": "2.0.2233",
4
4
  "description": "Canonical universal harness — AI-native software engineering via skill-driven orchestration; bootstraps plugkit for task execution and session isolation. Install in any AI coding agent host.",
5
5
  "author": "AnEntrypoint",
6
6
  "license": "MIT",