UNPKG

@blundergoat/goat-flow

Version:

AI coding agent harness and local dashboard for Claude Code, OpenAI Codex, Google Antigravity, and GitHub Copilot - setup audits, guardrails, structured skills, deny hooks, and persistent learning loops.

github.com/blundergoat/goat-flow

blundergoat/goat-flow

70 lines (59 loc) • 4.78 kB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
# Manifest snapshots Frozen copies of `workflow/manifest.json` at each release. Immutable after write. Used by CHANGELOG-scoped claim lint (M06b). ## Contract - **Live source of truth:** `workflow/manifest.json` (edited freely across milestones). - **Frozen snapshot:** `workflow/manifest-snapshots/vX.Y.Z.json` (never edited after release). - **Relationship:** at release time, a post-release script copies `workflow/manifest.json` into this directory under the release tag. The snapshot then drops out of the edit surface - all future changes land in the live file. ## Why snapshots exist CHANGELOG entries are frozen at release. The `## v1.1.0` section of `CHANGELOG.md` describes v1.1.0 as it shipped - "7 canonical skills", "12 setup + 4 agent build checks", "16 harness checks". When v1.2.3 changes a count, the v1.1.0 CHANGELOG section must still read the v1.1.0 numbers. Comparing CHANGELOG v1.1.0 claims against the live manifest would flag valid historical entries as drift; comparing against the matching snapshot catches only real edits to frozen text. ## Current snapshots | Version | File | Captures | |---------|------|----------| | v1.1.0 | `v1.1.0.json` | 7 skills, 12 setup + 4 agent + 16 harness = 32 checks, 7 dashboard views, 20 presets | | v1.2.0 | `v1.2.0.json` | 7 skills, 13 setup + 4 agent + 16 harness = 33 checks, 8 dashboard views, 22 presets | | v1.2.1 | `v1.2.1.json` | 7 skills, 13 setup + 4 agent + 16 harness = 33 checks, 8 dashboard views, 22 presets | | v1.2.2 | `v1.2.2.json` | 7 skills, 13 setup + 4 agent + 16 harness = 33 checks, 8 dashboard views, 22 presets | | v1.2.3 | `v1.2.3.json` | 7 skills, 13 setup + 4 agent + 16 harness = 33 checks, 8 dashboard views, 23 presets | | v1.2.4 | `v1.2.4.json` | 7 skills, 13 setup + 4 agent + 16 harness = 33 checks, 8 dashboard views, 28 presets | | v1.2.5 | `v1.2.5.json` | 7 skills, 13 setup + 4 agent + 16 harness = 33 checks, 8 dashboard views, 28 presets | | v1.3.0 | `v1.3.0.json` | 7 skills, 13 setup + 4 agent + 16 harness = 33 checks, 8 dashboard views, 29 presets | | v1.3.1 | `v1.3.1.json` | 7 skills, 13 setup + 4 agent + 16 harness = 33 checks, 8 dashboard views, 29 presets | | v1.3.2 | `v1.3.2.json` | 7 skills, 13 setup + 4 agent + 16 harness = 33 checks, 8 dashboard views, 29 presets | | v1.4.0 | `v1.4.0.json` | 7 skills, 13 setup + 4 agent + 17 harness = 34 checks, 8 dashboard views, 26 presets | | v1.4.1 | `v1.4.1.json` | 7 skills, 14 setup + 4 agent + 16 harness = 34 checks, 8 dashboard views, 26 presets | | v1.4.2 | `v1.4.2.json` | 7 skills, 14 setup + 4 agent + 17 harness = 35 checks, 8 dashboard views, 26 presets | | v1.4.3 | `v1.4.3.json` | 7 skills, 14 setup + 4 agent + 17 harness = 35 checks, 8 dashboard views, 26 presets | | v1.5.0 | `v1.5.0.json` | 7 skills, 14 setup + 4 agent + 16 harness = 34 checks, 8 dashboard views, 26 presets | | v1.5.1 | `v1.5.1.json` | 7 skills, 14 setup + 4 agent + 16 harness = 34 checks, 8 dashboard views, 26 presets | | v1.6.0 | `v1.6.0.json` | 7 skills, 14 setup + 4 agent + 16 harness = 34 checks, 9 dashboard views, 26 presets | | v1.6.1 | `v1.6.1.json` | 7 skills, 14 setup + 4 agent + 16 harness = 34 checks, 9 dashboard views, 26 presets | | v1.6.2 | `v1.6.2.json` | 7 skills, 14 setup + 4 agent + 16 harness = 34 checks, 9 dashboard views, 26 presets | | v1.6.3 | `v1.6.3.json` | 7 skills, 14 setup + 4 agent + 16 harness = 34 checks, 9 dashboard views, 26 presets | | v1.7.0 | `v1.7.0.json` | 7 skills, 15 setup + 4 agent + 17 harness = 36 checks, 11 dashboard views, 26 presets | | v1.8.0 | `v1.8.0.json` | 7 skills, 15 setup + 4 agent + 17 harness = 36 checks, 12 dashboard views, 26 presets | | v1.9.0 | `v1.9.0.json` | 7 skills, 15 setup + 4 agent + 17 harness = 36 checks, 12 dashboard views, 26 presets | | v1.9.1 | `v1.9.1.json` | 7 skills, 15 setup + 4 agent + 17 harness = 36 checks, 12 dashboard views, 26 presets | | v1.10.0 | `v1.10.0.json` | 7 skills, 15 setup + 4 agent + 17 harness = 36 checks, 12 dashboard views, 26 presets | | v1.10.1 | `v1.10.1.json` | 7 skills, 16 setup + 4 agent + 17 harness = 37 checks, 12 dashboard views, 26 presets | | v1.12.1 | `v1.12.1.json` | 7 skills, 16 setup + 4 agent + 17 harness = 37 checks, 12 dashboard views, 26 presets | ## Adding a snapshot at release time 1. `cp workflow/manifest.json workflow/manifest-snapshots/v${VERSION}.json` 2. Open the snapshot file, add the `_snapshot_note` and `snapshot_facts` blocks (mirror v1.1.0.json's shape). 3. Commit together with the release. M06b kill criterion (from the milestone): "If snapshots become burdensome, drop and accept that CHANGELOG lint only compares claims within the current version." A decision record in `.goat-flow/learning-loop/decisions/` captures that kill-switch when pulled.