sdg-agents 1.0.11 → 1.1.2
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
|
@@ -11,6 +11,12 @@
|
|
|
11
11
|
|
|
12
12
|
<br>
|
|
13
13
|
|
|
14
|
+
<p align="center">
|
|
15
|
+
Read the manifesto and visual guide at <a href="https://specdrivenguide.org">specdrivenguide.org</a>
|
|
16
|
+
</p>
|
|
17
|
+
|
|
18
|
+
<br>
|
|
19
|
+
|
|
14
20
|
<p align="center">
|
|
15
21
|
<kbd><img src="https://raw.githubusercontent.com/thiagocajadev/sgd-agents-cli/main/src/assets/img/sdg-agents-menu-v1.png" alt="Spec Driven Guide CLI in action" /></kbd>
|
|
16
22
|
</p>
|
|
@@ -83,22 +89,25 @@ Agent-specific entry files (`CLAUDE.md`, `.cursorrules`, `.windsurfrules`, etc.)
|
|
|
83
89
|
|
|
84
90
|
When you prefix a message to the agent, it enters the corresponding cycle:
|
|
85
91
|
|
|
86
|
-
| Trigger | Cycle | What happens
|
|
87
|
-
| :-------------------- | :------ |
|
|
88
|
-
| `land: <description>` | Land | Agent turns a raw vision into a grounded backlog of sequenced `feat:` tasks — runs before any code is written
|
|
89
|
-
| `feat: <description>` | Feature | Agent runs SPEC → PLAN → CODE → TEST → END
|
|
90
|
-
| `fix: <description>` | Fix | Agent runs SPEC → PLAN → CODE → TEST → END with RCA focus
|
|
91
|
-
| `docs: <description>` | Docs | Agent updates changelogs, ADRs, or specs
|
|
92
|
-
|
|
|
92
|
+
| Trigger | Cycle | What happens |
|
|
93
|
+
| :-------------------- | :------ | :-------------------------------------------------------------------------------------------------------------------------------------------------- |
|
|
94
|
+
| `land: <description>` | Land | Agent turns a raw vision into a grounded backlog of sequenced `feat:` tasks — runs before any code is written |
|
|
95
|
+
| `feat: <description>` | Feature | Agent runs SPEC → PLAN → CODE → TEST → END |
|
|
96
|
+
| `fix: <description>` | Fix | Agent runs SPEC → PLAN → CODE → TEST → END with RCA focus |
|
|
97
|
+
| `docs: <description>` | Docs | Agent updates changelogs, ADRs, or specs |
|
|
98
|
+
| `end:` | — | Close the active cycle — runs the END Phase checklist (changelog, backlog, commit). Also recovers a cycle if the agent loses track mid-conversation |
|
|
99
|
+
| No prefix | — | Agent asks: "land, feat, fix, or docs?" — then proceeds |
|
|
93
100
|
|
|
94
101
|
The agent **stops and waits for your approval** at SPEC and PLAN before writing any code.
|
|
95
102
|
|
|
96
103
|
```
|
|
97
104
|
SPEC → PLAN → CODE → TEST → END
|
|
98
|
-
↑ ↑
|
|
99
|
-
Wait Wait
|
|
105
|
+
↑ ↑ ↑
|
|
106
|
+
Wait Wait "end:"
|
|
100
107
|
```
|
|
101
108
|
|
|
109
|
+
> Type `end:` to close the active cycle. The agent runs the full END checklist — changelog, backlog sync, commit proposal. If the agent loses track mid-conversation, `end:` also recovers the cycle.
|
|
110
|
+
|
|
102
111
|
For a detailed walkthrough of each phase and its rules, see [Spec-Driven Development Guide](src/assets/dev-guides/spec-driven-dev-guide.md).
|
|
103
112
|
For a visual breakdown of the internal decision gates and loops, see [Agent Deep-Flow](src/assets/dev-guides/agent-deep-flow.md).
|
|
104
113
|
|
package/package.json
CHANGED
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "sdg-agents",
|
|
3
|
-
"version": "1.
|
|
3
|
+
"version": "1.1.2",
|
|
4
4
|
"description": "Structured working protocol and engineering rules for AI agents. Works with Claude Code, Antigravity, Codex, and others.",
|
|
5
5
|
"type": "module",
|
|
6
6
|
"main": "./src/engine/bin/index.mjs",
|
|
@@ -0,0 +1,67 @@
|
|
|
1
|
+
# Cycle Terminator — END Phase
|
|
2
|
+
|
|
3
|
+
This command has no arguments. Type `end:` to close the active cycle.
|
|
4
|
+
|
|
5
|
+
Executing the **Phase: END** checklist from the **Working Protocol** in strict order. Each item must be completed before advancing to the next.
|
|
6
|
+
|
|
7
|
+
---
|
|
8
|
+
|
|
9
|
+
## Step 1 — SUMMARIZE
|
|
10
|
+
|
|
11
|
+
Write one sentence per completed PLAN task. If no PLAN existed (e.g. `[S]` tasks), write a one-line summary of what changed.
|
|
12
|
+
|
|
13
|
+
## Step 2 — CHANGELOG
|
|
14
|
+
|
|
15
|
+
Append an entry under `## [Unreleased]` in `CHANGELOG.md`:
|
|
16
|
+
|
|
17
|
+
- `feat:` cycle → `### Added`
|
|
18
|
+
- `fix:` cycle → `### Fixed`
|
|
19
|
+
- `docs:` cycle → skip this step
|
|
20
|
+
- `land:` cycle → skip this step
|
|
21
|
+
|
|
22
|
+
If `## [Unreleased]` does not exist, create it above the most recent versioned entry.
|
|
23
|
+
|
|
24
|
+
## Step 3 — BACKLOG: tasks.md
|
|
25
|
+
|
|
26
|
+
- Move all `[DONE]` tasks to `## Done`.
|
|
27
|
+
- If no active `tasks.md` existed (`[S]` cycle), skip.
|
|
28
|
+
|
|
29
|
+
## Step 4 — BACKLOG: context.md
|
|
30
|
+
|
|
31
|
+
Update `## Now` with the next objective, or set it to `Ready for next instruction.` if nothing is pending.
|
|
32
|
+
|
|
33
|
+
## Step 5 — INSIGHTS
|
|
34
|
+
|
|
35
|
+
Log any patterns, findings, or rework discovered during this cycle in `context.md ## Engineering Insights`. Curate stale or irrelevant entries.
|
|
36
|
+
|
|
37
|
+
## Step 6 — CURATE
|
|
38
|
+
|
|
39
|
+
Scan all changed files for:
|
|
40
|
+
|
|
41
|
+
- Unfinished comments (`TODO`, `FIXME`, `...`)
|
|
42
|
+
- "AI-isms" or promotional language in docs/UI text
|
|
43
|
+
- Dead code introduced during the cycle
|
|
44
|
+
|
|
45
|
+
Run `git status` — confirm only intended files are staged.
|
|
46
|
+
|
|
47
|
+
## Step 7 — LINT
|
|
48
|
+
|
|
49
|
+
If a lint script exists (`lint`, `lint:fix`, `lint:all`, or a config file is detected):
|
|
50
|
+
|
|
51
|
+
- Run it and auto-fix what's possible.
|
|
52
|
+
- If non-auto-fixable violations remain, surface them explicitly.
|
|
53
|
+
- Block commit if errors remain.
|
|
54
|
+
|
|
55
|
+
## Step 8 — COMMIT
|
|
56
|
+
|
|
57
|
+
Propose the commit message and **WAIT** for explicit Developer approval before committing.
|
|
58
|
+
|
|
59
|
+
## Step 9 — PUSH
|
|
60
|
+
|
|
61
|
+
**ASK** for explicit permission before pushing to remote.
|
|
62
|
+
|
|
63
|
+
---
|
|
64
|
+
|
|
65
|
+
> [!WARNING]
|
|
66
|
+
> The cycle is **INCOMPLETE** until all applicable steps above are checked.
|
|
67
|
+
> Do not accept new work until END is fully executed.
|
|
@@ -9,13 +9,14 @@
|
|
|
9
9
|
|
|
10
10
|
On every request, classify intent before acting:
|
|
11
11
|
|
|
12
|
-
| Signal | Cycle
|
|
13
|
-
| :------------------------ |
|
|
14
|
-
| `land: ...` | Read `.ai/commands/sdg-land.md` → Follow **Land Cycle** (inception only)
|
|
15
|
-
| `feat: ...` | Read `.ai/commands/sdg-feat.md` (Context Charge) → Follow **Feature Cycle**
|
|
16
|
-
| `fix: ...` | Read `.ai/commands/sdg-fix.md` (Context Charge) → Follow **Fix Cycle**
|
|
17
|
-
| `docs: ...` | Read `.ai/commands/sdg-docs.md` (Context Charge) → Follow **Docs Cycle**
|
|
18
|
-
|
|
|
12
|
+
| Signal | Cycle |
|
|
13
|
+
| :------------------------ | :----------------------------------------------------------------------------- |
|
|
14
|
+
| `land: ...` | Read `.ai/commands/sdg-land.md` → Follow **Land Cycle** (inception only) |
|
|
15
|
+
| `feat: ...` | Read `.ai/commands/sdg-feat.md` (Context Charge) → Follow **Feature Cycle** |
|
|
16
|
+
| `fix: ...` | Read `.ai/commands/sdg-fix.md` (Context Charge) → Follow **Fix Cycle** |
|
|
17
|
+
| `docs: ...` | Read `.ai/commands/sdg-docs.md` (Context Charge) → Follow **Docs Cycle** |
|
|
18
|
+
| `end:` | Read `.ai/commands/sdg-end.md` → Execute **END Phase** checklist (no argument) |
|
|
19
|
+
| No prefix, intent unclear | Ask once: "land, feat, fix, or docs?" — then proceed |
|
|
19
20
|
|
|
20
21
|
---
|
|
21
22
|
|
|
@@ -254,7 +254,8 @@ function buildMasterInstructions(selections) {
|
|
|
254
254
|
| \`.ai/commands/sdg-land.md\` | Land Cycle (Project Inception & Backlog) |
|
|
255
255
|
| \`.ai/commands/sdg-feat.md\` | Feature Cycle (Specs & Implementation) |
|
|
256
256
|
| \`.ai/commands/sdg-fix.md\` | Fix Cycle (Forensics & Regression) |
|
|
257
|
-
| \`.ai/commands/sdg-docs.md\` | Documentation Cycle (ADRs & Logs)
|
|
257
|
+
| \`.ai/commands/sdg-docs.md\` | Documentation Cycle (ADRs & Logs) |
|
|
258
|
+
| \`.ai/commands/sdg-end.md\` | END Phase — close the active cycle (changelog, backlog, commit) |`;
|
|
258
259
|
|
|
259
260
|
return workingCyclesString;
|
|
260
261
|
}
|
|
@@ -377,6 +378,7 @@ function buildClaudeContent() {
|
|
|
377
378
|
| \`feat: ...\` | Feature Cycle — read \`.ai/commands/sdg-feat.md\` |
|
|
378
379
|
| \`fix: ...\` | Fix Cycle — read \`.ai/commands/sdg-fix.md\` |
|
|
379
380
|
| \`docs: ...\` | Docs Cycle — read \`.ai/commands/sdg-docs.md\` |
|
|
381
|
+
| \`end:\` | END Phase — read \`.ai/commands/sdg-end.md\` |
|
|
380
382
|
`;
|
|
381
383
|
|
|
382
384
|
return claudeContent;
|