@pieai/pro-gov 0.3.3 → 0.3.4
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 +80 -15
- package/assets/docs/reference/adoption/adoption-playbook.md +42 -6
- package/assets/docs/reference/adoption/downstream-project-registry.md +24 -17
- package/assets/docs/reference/adoption/project-relationship.md +9 -8
- package/assets/docs/reference/adoption/public-release-checklist.md +34 -29
- package/assets/docs/reference/adoption/recommended-agent-tooling.md +106 -0
- package/assets/docs/reference/adoption/site-publication-brief.md +22 -14
- package/assets/integrations/ponytail.md +109 -0
- package/assets/integrations/superpowers.md +30 -2
- package/assets/starter/.gemini/settings.json +5 -0
- package/assets/starter/.github/workflows/docs-check.yml +4 -2
- package/assets/starter/docs/governance/agents-routing/doc-only-v0.9.md +3 -2
- package/assets/starter/docs/governance/agents-routing/engineering-runtime-v0.9.md +3 -2
- package/assets/starter/docs/governance/ssot-v0.9.md +2 -2
- package/assets/starter/lefthook.template.yml +2 -2
- package/cli-guide.md +25 -1
- package/dist/cli.js +1444 -18
- package/package.json +11 -10
|
@@ -0,0 +1,109 @@
|
|
|
1
|
+
# Ponytail Integration
|
|
2
|
+
|
|
3
|
+
Ponytail is an external plugin that asks an AI to prefer the smallest solution
|
|
4
|
+
that still works. Project Governance System does not vendor, rewrite, or require
|
|
5
|
+
the Ponytail plugin.
|
|
6
|
+
|
|
7
|
+
## The Beginner Version
|
|
8
|
+
|
|
9
|
+
Think of an AI project as a building site:
|
|
10
|
+
|
|
11
|
+
- PGS is the traffic desk and inspection station.
|
|
12
|
+
- Superpowers is the construction process.
|
|
13
|
+
- Ponytail is the cost and complexity adviser.
|
|
14
|
+
|
|
15
|
+
The adviser can stop the team from buying unnecessary materials. The adviser
|
|
16
|
+
cannot cancel the fire exit, the safety inspection, or a room the owner
|
|
17
|
+
explicitly requested.
|
|
18
|
+
|
|
19
|
+
## Safe Default
|
|
20
|
+
|
|
21
|
+
Keep the global Ponytail mode `off`.
|
|
22
|
+
|
|
23
|
+
`off` does not mean Ponytail is useless. It means Ponytail is available when a
|
|
24
|
+
task needs a simplicity review without silently influencing every project and
|
|
25
|
+
every answer.
|
|
26
|
+
|
|
27
|
+
Do not describe `lite` as "almost off." In Ponytail 4.7.0, `lite` still injects
|
|
28
|
+
the shared minimalism rules into the active AI context. Its mode-specific line
|
|
29
|
+
is gentler than `full`, but the same AI still sees rules about fewer files,
|
|
30
|
+
fewer abstractions, shorter diffs, and avoiding speculative work.
|
|
31
|
+
|
|
32
|
+
## Mode Policy
|
|
33
|
+
|
|
34
|
+
| Mode | Governed use |
|
|
35
|
+
| --- | --- |
|
|
36
|
+
| `off` | Recommended global default. Ponytail does not inject its active ruleset. |
|
|
37
|
+
| `lite` | First mode to test in one isolated task or session. Complete the requested work, while surfacing a simpler option. |
|
|
38
|
+
| `full` | Optional stress test after `lite`. Use only in an isolated task to see whether stronger minimalism harms scope or proof. |
|
|
39
|
+
| `ultra` | Not part of the recommended PGS workflow. It is too aggressive for a default governed engineering lane. |
|
|
40
|
+
|
|
41
|
+
Return to `off` after every comparison.
|
|
42
|
+
|
|
43
|
+
## Comparison Protocol
|
|
44
|
+
|
|
45
|
+
Use the same bounded, low-risk task in separate sessions or worktrees:
|
|
46
|
+
|
|
47
|
+
```text
|
|
48
|
+
baseline with off
|
|
49
|
+
-> repeat with lite
|
|
50
|
+
-> compare scope, proof, complexity, and clarity
|
|
51
|
+
-> optionally repeat with full
|
|
52
|
+
-> return global mode to off
|
|
53
|
+
```
|
|
54
|
+
|
|
55
|
+
Compare:
|
|
56
|
+
|
|
57
|
+
- whether every requested requirement was delivered;
|
|
58
|
+
- whether Superpowers brainstorming, planning, TDD, debugging, and verification
|
|
59
|
+
gates were preserved;
|
|
60
|
+
- whether tests, security, accessibility, validation, and data-loss prevention
|
|
61
|
+
remained intact;
|
|
62
|
+
- files, dependencies, abstractions, and lines changed;
|
|
63
|
+
- whether the explanation and durable evidence stayed understandable;
|
|
64
|
+
- token and time measurements only when the AI host exposes trustworthy data.
|
|
65
|
+
|
|
66
|
+
A shorter answer or smaller diff is not automatically a better result.
|
|
67
|
+
|
|
68
|
+
## Priority And Boundary
|
|
69
|
+
|
|
70
|
+
Use this priority order:
|
|
71
|
+
|
|
72
|
+
1. User instructions and project safety requirements.
|
|
73
|
+
2. PGS routing, governance boundaries, and required evidence.
|
|
74
|
+
3. Superpowers workflow gates.
|
|
75
|
+
4. Ponytail complexity and cost advice.
|
|
76
|
+
|
|
77
|
+
Ponytail may:
|
|
78
|
+
|
|
79
|
+
- question speculative scope;
|
|
80
|
+
- prefer the standard library or an already-installed dependency;
|
|
81
|
+
- suggest fewer files, dependencies, and abstractions;
|
|
82
|
+
- identify code or structure that can be removed safely.
|
|
83
|
+
|
|
84
|
+
Ponytail may not:
|
|
85
|
+
|
|
86
|
+
- remove an explicit requirement;
|
|
87
|
+
- bypass PGS routing or document placement;
|
|
88
|
+
- skip required brainstorming, plans, TDD, debugging, or verification;
|
|
89
|
+
- weaken security, accessibility, trust-boundary validation, error handling, or
|
|
90
|
+
data-loss protection;
|
|
91
|
+
- replace proof with "the code is shorter";
|
|
92
|
+
- silently change the user's global Ponytail configuration.
|
|
93
|
+
|
|
94
|
+
## Recommended Workflow
|
|
95
|
+
|
|
96
|
+
```mermaid
|
|
97
|
+
flowchart TD
|
|
98
|
+
A["Task arrives"] --> B["PGS selects the project lane"]
|
|
99
|
+
B --> C["Superpowers runs the required workflow"]
|
|
100
|
+
C --> D{"Would a simplicity review help?"}
|
|
101
|
+
D -- "no" --> E["Implement and verify normally"]
|
|
102
|
+
D -- "yes" --> F["Use Ponytail explicitly for this task"]
|
|
103
|
+
F --> G["Keep requirements and proof intact"]
|
|
104
|
+
G --> E
|
|
105
|
+
```
|
|
106
|
+
|
|
107
|
+
For normal engineering, Ponytail is a scoped adviser, not a second workflow
|
|
108
|
+
engine and not a replacement for Superpowers.
|
|
109
|
+
|
|
@@ -2,6 +2,18 @@
|
|
|
2
2
|
|
|
3
3
|
Superpowers is an external plugin/system. This repository does not vendor or rewrite it.
|
|
4
4
|
|
|
5
|
+
## The Beginner Version
|
|
6
|
+
|
|
7
|
+
Think of an AI project as a building site:
|
|
8
|
+
|
|
9
|
+
- PGS is the traffic desk and inspection station.
|
|
10
|
+
- Superpowers is the construction process.
|
|
11
|
+
- Ponytail is an optional cost and complexity adviser.
|
|
12
|
+
|
|
13
|
+
PGS chooses the lane. Superpowers makes sure the work follows the right
|
|
14
|
+
engineering process. Ponytail may suggest a leaner implementation, but it
|
|
15
|
+
cannot cancel the process or the final inspection.
|
|
16
|
+
|
|
5
17
|
## Boundary
|
|
6
18
|
|
|
7
19
|
Superpowers owns engineering workflows such as:
|
|
@@ -20,6 +32,13 @@ Project Governance System owns:
|
|
|
20
32
|
- current work index conventions
|
|
21
33
|
- the governed location and boundary for externally sourced AI evidence rules
|
|
22
34
|
|
|
35
|
+
Ponytail, when explicitly enabled for a task, may advise on:
|
|
36
|
+
|
|
37
|
+
- speculative scope;
|
|
38
|
+
- unnecessary dependencies;
|
|
39
|
+
- avoidable files and abstractions;
|
|
40
|
+
- simpler implementation choices.
|
|
41
|
+
|
|
23
42
|
## Rule
|
|
24
43
|
|
|
25
44
|
Use Superpowers inside the selected project lane. Do not let Superpowers create a separate durable document tree unless the project explicitly adopts one.
|
|
@@ -33,6 +52,9 @@ Durable outputs should map back to the project's doc-gov layers:
|
|
|
33
52
|
## Execution Order
|
|
34
53
|
|
|
35
54
|
Agents routing classifies first. Superpowers executes inside the selected lane.
|
|
55
|
+
Optional Ponytail advice comes after those responsibilities are known. It may
|
|
56
|
+
make the implementation leaner, but it must not remove explicit requirements or
|
|
57
|
+
reduce how correctness is proven.
|
|
36
58
|
|
|
37
59
|
```mermaid
|
|
38
60
|
flowchart TD
|
|
@@ -41,8 +63,11 @@ flowchart TD
|
|
|
41
63
|
C --> D{"Does this lane need a Superpowers workflow?"}
|
|
42
64
|
D -- "yes" --> E["Use the matching Superpowers skill"]
|
|
43
65
|
D -- "no" --> F["Use the local non-Superpowers lane rules"]
|
|
44
|
-
E --> G
|
|
45
|
-
F -->
|
|
66
|
+
E --> G{"Would an explicit simplicity review help?"}
|
|
67
|
+
F --> H["Write durable outputs into doc-gov layers"]
|
|
68
|
+
G -- "yes" --> I["Use Ponytail without dropping workflow gates"]
|
|
69
|
+
G -- "no" --> H
|
|
70
|
+
I --> H
|
|
46
71
|
```
|
|
47
72
|
|
|
48
73
|
If Superpowers suggests a default location such as `docs/superpowers/**`, project instructions may override that location. The durable project record should still land in the governed doc-gov layer unless the project has explicitly adopted a separate Superpowers document tree.
|
|
@@ -50,3 +75,6 @@ If Superpowers suggests a default location such as `docs/superpowers/**`, projec
|
|
|
50
75
|
Host-specific files such as `CLAUDE.md` may include Superpowers skill routing
|
|
51
76
|
text. That text is an adapter. It must not replace the project `AGENTS.md`
|
|
52
77
|
router or run before the Project Governance System routing block.
|
|
78
|
+
|
|
79
|
+
For Ponytail mode policy and the isolated comparison protocol, read
|
|
80
|
+
`integrations/ponytail.md`.
|
|
@@ -5,7 +5,7 @@ on:
|
|
|
5
5
|
paths:
|
|
6
6
|
- "AGENTS.md"
|
|
7
7
|
- "CLAUDE.md"
|
|
8
|
-
- "
|
|
8
|
+
- ".gemini/settings.json"
|
|
9
9
|
- "README.md"
|
|
10
10
|
- "docs/**"
|
|
11
11
|
- "lefthook.yml"
|
|
@@ -18,7 +18,7 @@ on:
|
|
|
18
18
|
paths:
|
|
19
19
|
- "AGENTS.md"
|
|
20
20
|
- "CLAUDE.md"
|
|
21
|
-
- "
|
|
21
|
+
- ".gemini/settings.json"
|
|
22
22
|
- "README.md"
|
|
23
23
|
- "docs/**"
|
|
24
24
|
- "lefthook.yml"
|
|
@@ -34,6 +34,8 @@ jobs:
|
|
|
34
34
|
- uses: actions/checkout@v4
|
|
35
35
|
|
|
36
36
|
- uses: pnpm/action-setup@v4
|
|
37
|
+
with:
|
|
38
|
+
version: 10.26.0
|
|
37
39
|
|
|
38
40
|
- uses: actions/setup-node@v4
|
|
39
41
|
with:
|
|
@@ -79,5 +79,6 @@ unless the project explicitly opts them into doc-gov.
|
|
|
79
79
|
## External Workflow Boundary
|
|
80
80
|
|
|
81
81
|
This route runs before external workflow systems such as Superpowers or GStack.
|
|
82
|
-
Host-specific
|
|
83
|
-
specific AI client, but they must not replace the project
|
|
82
|
+
Host-specific adapters such as `CLAUDE.md` or `.gemini/settings.json` may adapt
|
|
83
|
+
the route for a specific AI client, but they must not replace the project
|
|
84
|
+
`AGENTS.md` route.
|
|
@@ -74,5 +74,6 @@ This route runs before external workflow systems such as Superpowers or GStack.
|
|
|
74
74
|
Those systems may provide skills, reviews, browser workflows, or shipping gates,
|
|
75
75
|
but they execute **inside** the lane selected by this route.
|
|
76
76
|
|
|
77
|
-
Host-specific
|
|
78
|
-
specific AI client, but they must not replace the project
|
|
77
|
+
Host-specific adapters such as `CLAUDE.md` or `.gemini/settings.json` may adapt
|
|
78
|
+
the route for a specific AI client, but they must not replace the project
|
|
79
|
+
`AGENTS.md` route.
|
|
@@ -71,7 +71,7 @@ Extra governed roots are allowed only when a project explicitly opts in.
|
|
|
71
71
|
|
|
72
72
|
Before changing docs, look for project-local guidance in this order:
|
|
73
73
|
|
|
74
|
-
1. `AGENTS.md`, `CLAUDE.md`,
|
|
74
|
+
1. `AGENTS.md`, `CLAUDE.md`, `.gemini/settings.json`, or equivalent AI router/config adapter.
|
|
75
75
|
2. `docs/governance/boundary.md`.
|
|
76
76
|
3. `docs/governance/ssot-v0.9.md`.
|
|
77
77
|
4. `docs/governance/agents-routing/` and the project's selected agents-routing file.
|
|
@@ -100,7 +100,7 @@ works for governed docs:
|
|
|
100
100
|
|
|
101
101
|
| Need | Usually belongs in |
|
|
102
102
|
| --- | --- |
|
|
103
|
-
| AI entry and startup routing | `AGENTS.md` plus host-specific
|
|
103
|
+
| AI entry and startup routing | `AGENTS.md` plus thin host-specific adapters such as `CLAUDE.md` and `.gemini/settings.json` |
|
|
104
104
|
| Agents-routing rules | `docs/governance/agents-routing/` |
|
|
105
105
|
| Doc-system rules, templates, and manifest | `docs/governance/` |
|
|
106
106
|
| Project AI/development policy | `docs/policy/` |
|
|
@@ -6,10 +6,10 @@ pre-commit:
|
|
|
6
6
|
parallel: false
|
|
7
7
|
commands:
|
|
8
8
|
01-doc-gov-router-check:
|
|
9
|
-
glob: "{AGENTS.md,CLAUDE.md,
|
|
9
|
+
glob: "{AGENTS.md,CLAUDE.md,README.md,.gemini/settings.json,docs/**/*.{md,yml,yaml}}"
|
|
10
10
|
run: pnpm doc-gov router-check
|
|
11
11
|
02-doc-gov-check:
|
|
12
|
-
glob: "{docs/**/*.md,docs/governance/**/*.{md,yml,yaml},docs/policy/**/*.md,AGENTS.md,CLAUDE.md
|
|
12
|
+
glob: "{docs/**/*.md,docs/governance/**/*.{md,yml,yaml},docs/policy/**/*.md,AGENTS.md,CLAUDE.md,.gemini/settings.json}"
|
|
13
13
|
run: pnpm doc-gov check && pnpm doc-gov scan --check && pnpm doc-gov links && pnpm doc-gov audit
|
|
14
14
|
|
|
15
15
|
commit-msg:
|
package/cli-guide.md
CHANGED
|
@@ -3,14 +3,21 @@
|
|
|
3
3
|
`pro-gov` is the project-level companion to `doc-gov`.
|
|
4
4
|
|
|
5
5
|
Use it to inspect packaged Project Governance System assets, plan a starter
|
|
6
|
-
installation, compare reusable governance files,
|
|
6
|
+
installation, compare reusable governance files, inspect local projects, and run
|
|
7
|
+
package health checks.
|
|
7
8
|
|
|
8
9
|
The validator remains `doc-gov`.
|
|
9
10
|
|
|
10
11
|
## Commands
|
|
11
12
|
|
|
13
|
+
Public package-safe commands:
|
|
14
|
+
|
|
12
15
|
```bash
|
|
13
16
|
pro-gov assets list
|
|
17
|
+
pro-gov assets discover --target .
|
|
18
|
+
pro-gov assets recommend --target .
|
|
19
|
+
pro-gov lens inspect --target .
|
|
20
|
+
pro-gov lens report --target . --out .pro-gov/lens-report.md
|
|
14
21
|
pro-gov init --profile engineering-runtime --dry-run
|
|
15
22
|
pro-gov init --profile doc-only --dry-run
|
|
16
23
|
pro-gov sync --check
|
|
@@ -20,11 +27,28 @@ pro-gov doctor
|
|
|
20
27
|
`init` and `sync` are read-only in the first release. They report planned files,
|
|
21
28
|
missing files, or changed files, but they do not overwrite target projects.
|
|
22
29
|
|
|
30
|
+
Full upstream-checkout commands:
|
|
31
|
+
|
|
32
|
+
```bash
|
|
33
|
+
pro-gov assets list --json
|
|
34
|
+
pro-gov assets plan --bundle base-governance --target . --out .pro-gov/asset-plan.json
|
|
35
|
+
pro-gov assets apply --plan .pro-gov/asset-plan.json
|
|
36
|
+
pro-gov assets check --target .
|
|
37
|
+
pro-gov assets npx add <source> --plan
|
|
38
|
+
pro-gov assets npx update --plan
|
|
39
|
+
```
|
|
40
|
+
|
|
41
|
+
These commands need the private `agent-assets/` registry in a full Project
|
|
42
|
+
Governance System checkout. The public npm package excludes private and
|
|
43
|
+
third-party skill bodies by design.
|
|
44
|
+
|
|
23
45
|
## Typical Adoption Flow
|
|
24
46
|
|
|
25
47
|
```bash
|
|
26
48
|
pnpm add -D @pieai/pro-gov @pieai/doc-gov
|
|
27
49
|
pnpm pro-gov assets list
|
|
50
|
+
pnpm pro-gov assets discover --target .
|
|
51
|
+
pnpm pro-gov assets recommend --target .
|
|
28
52
|
pnpm pro-gov init --profile engineering-runtime --dry-run
|
|
29
53
|
pnpm pro-gov sync --check
|
|
30
54
|
pnpm doc-gov migrate --profile engineering-runtime --check
|