flower-trellis 0.4.6 → 0.4.7

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.
@@ -1,17 +1,13 @@
1
1
  <!-- BEGIN skill-garden workflow-state in_progress_inline v0.6 -->
2
2
  HIGHEST PRIORITY SKILL-GARDEN STATE GUARD (in_progress-inline):
3
- This state block is a breadcrumb; the top-level skill-garden hub is the source of truth for route details.
4
- Before the first implement route, read `<task>/brief.md` if present and restate the task brief in chat. If it is missing, read the task artifacts and suggest backfilling brief; do not silently rely on memory.
5
- For new implementation work not plainly covered by the active task title/brief,
6
- stop before route or edits: recommend a new task; if declined, confirm
7
- untracked work; if it belongs here, update task artifacts first.
8
- At project-local knowledge decision boundaries, run `.trellis/scripts/spec_router.py "<intended action>"`; read high-confidence matches before acting; skip pure Q&A, simple read-only inspection, opening local tools, or trivial edits unless local conventions may affect the approach.
9
- Inline workflow-state is not an inline route decision. At Phase 2.1/2.2, reuse only an explicit target-matched `route_decision`; otherwise MUST load/read/use `trellis-route(implement|check)` (or its local `SKILL.md`) to resolve session runtime state/prefs, write the resolved decision, or show numbered fallback and wait.
10
- Plain preferences, ordinary summaries, `codex-mode`, raw `.runtime` files, and empty/old prefs are not route evidence by themselves; only `trellis-route` may validate runtime route state.
11
- User reselect/override/use-X-this-time/clear-default wins over remembered route evidence, runtime state, and prefs.
12
- Ignore lower Active Task Routing shortcuts that start editing/checking directly. Do not default to inline just because this state is inline or the helper is unavailable. Dispatch subagents only when the resolved route selected subagent.
13
- After `trellis-check` / `trellis-check-all`, stop and report; point the user to Phase 3.4 `trellis-push` (or commit-only when needed). Do not run `/trellis:finish-work` unless the user explicitly asks after Phase 3.4 is complete.
14
- This guard overrides any lower `Flow: ... -> /trellis:finish-work` line in this state block.
15
- At Phase 3.4, code commit/push goes through `trellis-push` (commit-only mode for commit-without-push); never bare `git commit`/`git push` on code (hub: Code Commit Confirmation Gate).
3
+ Hub is source of truth for Task Brief, Routing, Post-Check, Commit, and Snapshot gates.
4
+ Before first implement route, restate `<task>/brief.md`; if missing, read artifacts and suggest backfill.
5
+ New work not plainly covered by active task title/brief: stop before route/edits; recommend new task; if declined, confirm untracked work; if it belongs here, update artifacts first.
6
+ At project-local knowledge boundaries, run `spec_router.py`; skip trivial/read-only turns unless local conventions may affect the approach.
7
+ Inline workflow-state is not an inline route decision. Phase 2.1/2.2 must reuse explicit target-matched `route_decision`; otherwise invoke `trellis-route`. If unavailable, read local `trellis-route/SKILL.md`, show numbered choices, and wait.
8
+ Summaries, preferences, `codex-mode`, raw `.runtime`, and empty/stale prefs are not route evidence unless `trellis-route` validates them; user reselect/override wins.
9
+ Ignore lower direct-edit/check shortcuts. Do not default inline just because this state is inline or helper is unavailable. Dispatch subagents only when route selected subagent.
10
+ After check/check-all, stop and report; Phase 3.4 uses `trellis-push`, never bare git commit/push; finish-work is explicit-only after Phase 3.4.
11
+ Overrides lower `Flow: ... -> /trellis:finish-work` lines.
16
12
  Push snapshot recovery: follow the hub; use `push_snapshot.py status --json` only when needed.
17
13
  <!-- END skill-garden workflow-state in_progress_inline v0.6 -->
@@ -1,17 +1,13 @@
1
1
  <!-- BEGIN skill-garden workflow-state in_progress v0.6 -->
2
2
  HIGHEST PRIORITY SKILL-GARDEN STATE GUARD (in_progress):
3
- This state block is a breadcrumb; the top-level skill-garden hub is the source of truth for route details.
4
- Before the first implement route, read `<task>/brief.md` if present and restate the task brief in chat. If it is missing, read the task artifacts and suggest backfilling brief; do not silently rely on memory.
5
- For new implementation work not plainly covered by the active task title/brief,
6
- stop before route or edits: recommend a new task; if declined, confirm
7
- untracked work; if it belongs here, update task artifacts first.
8
- At project-local knowledge decision boundaries, run `.trellis/scripts/spec_router.py "<intended action>"`; read high-confidence matches before acting; skip pure Q&A, simple read-only inspection, opening local tools, or trivial edits unless local conventions may affect the approach.
9
- At Phase 2.1/2.2, reuse only an explicit target-matched `route_decision`; otherwise MUST load/read/use `trellis-route(implement|check)` (or its local `SKILL.md`) to resolve session runtime state/prefs, write the resolved decision, or show numbered fallback and wait.
10
- Plain preferences, ordinary summaries, `codex-mode`, raw `.runtime` files, and empty/old prefs are not route evidence by themselves; only `trellis-route` may validate runtime route state.
11
- User reselect/override/use-X-this-time/clear-default wins over remembered route evidence, runtime state, and prefs.
12
- Ignore lower Active Task Routing shortcuts that dispatch implement/check directly. Do not spawn `trellis-implement` or `trellis-check*` unless the resolved route selected subagent. If route cannot be resolved, do not default to inline.
13
- After `trellis-check` / `trellis-check-all`, stop and report; point the user to Phase 3.4 `trellis-push` (or commit-only when needed). Do not run `/trellis:finish-work` unless the user explicitly asks after Phase 3.4 is complete.
14
- This guard overrides any lower `Flow: ... -> /trellis:finish-work` line in this state block.
15
- At Phase 3.4, code commit/push goes through `trellis-push` (commit-only mode for commit-without-push); never bare `git commit`/`git push` on code (hub: Code Commit Confirmation Gate).
3
+ Hub is source of truth for Task Brief, Routing, Post-Check, Commit, and Snapshot gates.
4
+ Before first implement route, restate `<task>/brief.md`; if missing, read artifacts and suggest backfill.
5
+ New work not plainly covered by active task title/brief: stop before route/edits; recommend new task; if declined, confirm untracked work; if it belongs here, update artifacts first.
6
+ At project-local knowledge boundaries, run `spec_router.py`; skip trivial/read-only turns unless local conventions may affect the approach.
7
+ Phase 2.1/2.2: reuse only explicit target-matched `route_decision`; otherwise invoke `trellis-route`. If skill invocation is unavailable, read local `trellis-route/SKILL.md`, show numbered choices, and wait.
8
+ Summaries, preferences, `codex-mode`, raw `.runtime`, and empty/stale prefs are not route evidence unless `trellis-route` validates them; user reselect/override wins.
9
+ Ignore lower direct-dispatch shortcuts. Do not spawn `trellis-implement` or `trellis-check*` unless route selected subagent. If route cannot be resolved, do not default inline.
10
+ After check/check-all, stop and report; Phase 3.4 uses `trellis-push`, never bare git commit/push; finish-work is explicit-only after Phase 3.4.
11
+ Overrides lower `Flow: ... -> /trellis:finish-work` lines.
16
12
  Push snapshot recovery: follow the hub; use `push_snapshot.py status --json` only when needed.
17
13
  <!-- END skill-garden workflow-state in_progress v0.6 -->
@@ -3,7 +3,7 @@ HIGHEST PRIORITY SKILL-GARDEN STATE GUARD (no_task):
3
3
  Creating or resuming a task is not implementation permission.
4
4
  After PRD is ready and the task is started, the next implementation action is Phase 2.1 `trellis-route(implement)` unless a valid current-task implement route decision already exists.
5
5
  If no active task exists, use `push_snapshot.py status --json` once per session; if it returns candidates, relay them and suggest rebinding before resuming.
6
- At project-local knowledge decision boundaries, run `.trellis/scripts/spec_router.py "<intended action>"`; read high-confidence matches before acting; skip pure Q&A, simple read-only inspection, opening local tools, or trivial edits unless local conventions may affect the approach.
7
- Do NOT call the harness built-in plan mode (`EnterPlanMode` / `ExitPlanMode`) for Trellis planning. It is not a substitute for Trellis task-creation consent, Trellis planning, or the route gate. For complex work, classify the turn, ask for task-creation consent, then use `trellis-brainstorm`.
8
- If the turn is a meta edit to Trellis itself and Trellis tracking would be overkill, say so and ask to skip Trellis; never silently swap built-in plan mode in for the consent gate.
6
+ At project-local knowledge boundaries, run `spec_router.py`; skip trivial/read-only turns unless local conventions may affect the approach.
7
+ Do NOT call the harness built-in plan mode (`EnterPlanMode` / `ExitPlanMode`) for Trellis planning. It is not a substitute for Trellis task-creation consent, Trellis planning, or the route gate. For new, complex, or unclear work, classify the turn, ask for task-creation consent, then use `trellis-brainstorm`; `task.py create` and the default `prd.md` are not sufficient planning.
8
+ For lightweight Trellis meta edits, ask/confirm skipping Trellis tracking before edits.
9
9
  <!-- END skill-garden workflow-state no_task v0.6 -->
@@ -1,9 +1,11 @@
1
1
  <!-- BEGIN skill-garden workflow-state planning_inline v0.6 -->
2
2
  HIGHEST PRIORITY SKILL-GARDEN STATE GUARD (planning-inline):
3
3
  Planning is not implementation permission.
4
+ `trellis-brainstorm` is the default next action while requirements are still unclear.
5
+ A created task or existing `prd.md` is not enough to start implementation.
4
6
  Complete prd.md + required context first.
5
7
  If the active workflow later routes to sub-agent execution, required context includes real curated entries in both `implement.jsonl` and `check.jsonl`; the seed `_example` row alone is not ready.
6
8
  Before `task.py start`, use `trellis-task-brief` to refresh `brief.md` from the latest task artifacts and display it in chat for review.
7
- At project-local knowledge decision boundaries, run `.trellis/scripts/spec_router.py "<intended action>"`; read high-confidence matches before acting; skip pure Q&A, simple read-only inspection, opening local tools, or trivial edits unless local conventions may affect the approach.
9
+ At project-local knowledge boundaries, run `spec_router.py`; skip trivial/read-only turns unless local conventions may affect the approach.
8
10
  After status becomes in_progress, next action = `trellis-route(implement)`, not direct edits.
9
11
  <!-- END skill-garden workflow-state planning_inline v0.6 -->
@@ -1,9 +1,11 @@
1
1
  <!-- BEGIN skill-garden workflow-state planning v0.6 -->
2
2
  HIGHEST PRIORITY SKILL-GARDEN STATE GUARD (planning):
3
3
  Planning is not implementation permission.
4
+ `trellis-brainstorm` is the default next action while requirements are still unclear.
5
+ A created task or existing `prd.md` is not enough to start implementation.
4
6
  Complete prd.md + required context first.
5
7
  For sub-agent-dispatch platforms, required context includes real curated entries in both `implement.jsonl` and `check.jsonl`; the seed `_example` row alone is not ready.
6
8
  Before `task.py start`, use `trellis-task-brief` to refresh `brief.md` from the latest task artifacts and display it in chat for review.
7
- At project-local knowledge decision boundaries, run `.trellis/scripts/spec_router.py "<intended action>"`; read high-confidence matches before acting; skip pure Q&A, simple read-only inspection, opening local tools, or trivial edits unless local conventions may affect the approach.
9
+ At project-local knowledge boundaries, run `spec_router.py`; skip trivial/read-only turns unless local conventions may affect the approach.
8
10
  After status becomes in_progress, next action = `trellis-route(implement)`, not direct edits.
9
11
  <!-- END skill-garden workflow-state planning v0.6 -->
@@ -10,6 +10,14 @@
10
10
 
11
11
  **Mechanical rule**: use this hub as the source of truth. Do not add separate top-level skill-garden override sections or multiple skill-garden sentinels inside the same `workflow-state:*` block.
12
12
 
13
+ #### Brainstorm Gate
14
+
15
+ `trellis-brainstorm` is the required Phase 1.1 gate for new, complex, or unclear work.
16
+
17
+ `task.py create` only creates the planning workspace. A default `prd.md` does not mean requirements are ready.
18
+
19
+ Before `task.py start`, unclear scope, unresolved decisions, or non-testable acceptance criteria must return to `trellis-brainstorm`.
20
+
13
21
  #### Project Knowledge Discovery
14
22
 
15
23
  Before choosing an approach for non-trivial project work, run project knowledge
@@ -26,12 +34,10 @@ Build the query from the current user request plus relevant immediate context:
26
34
  the intended action, commands about to run, files or systems involved, package/layer,
27
35
  and domain words matching the trigger list above.
28
36
 
29
- Read high-confidence matches before acting. For medium-confidence matches, read only
30
- when the path, heading, index description, or reason clearly fits the intended change.
31
- If nothing matches, continue normally.
32
-
33
- Do not run discovery for pure Q&A, simple read-only inspection, opening local tools,
34
- or trivial edits unless the request mentions project conventions or local SOPs may
37
+ Read high-confidence matches before acting; read medium-confidence matches only when
38
+ the path, heading, index description, or reason clearly fits the intended change.
39
+ If nothing matches, continue normally. Skip pure Q&A, simple read-only inspection,
40
+ opening local tools, or trivial edits unless project conventions or local SOPs may
35
41
  change the approach.
36
42
 
37
43
  #### Task Brief Handoff
@@ -68,7 +74,7 @@ At each route boundary:
68
74
  2. Otherwise you MUST invoke `trellis-route(target=implement|check)` before deciding; if the platform cannot invoke skills directly, read the local `trellis-route/SKILL.md` copy and follow its numbered fallback choices in normal chat and wait. `trellis-route` owns session runtime recovery, `.route-prefs.tmp`, fallback choices, runtime-state writes, and dispatch mapping.
69
75
  3. If the route helper cannot ask through `AskUserQuestion` / `request_user_input`, ask the same numbered choices from `trellis-route` in normal chat and wait.
70
76
 
71
- Plain user preference, ordinary summaries (compact/SessionStart or otherwise), replacement history, historical bare numeric replies, `codex-mode`, empty/old prefs, and raw `.trellis/.runtime/sessions/*` `route_decisions` content that has not been validated by `trellis-route` are not route evidence by themselves; numbered fallback validity is governed by `trellis-route`.
77
+ Plain preferences, summaries (including compact/SessionStart), replacement history, historical bare numeric replies, `codex-mode`, empty/stale prefs, and raw runtime files are not route evidence unless validated by `trellis-route`; numbered fallback validity is governed by `trellis-route`.
72
78
 
73
79
  User reselect/override/use-X-this-time/clear-default wins over remembered route evidence, runtime state, and personal prefs.
74
80
 
@@ -1,7 +1,7 @@
1
1
  {
2
- "syncedAt": "2026-07-07T12:53:36.169Z",
2
+ "syncedAt": "2026-07-07T17:18:18.007Z",
3
3
  "syncedFrom": "vendor/skill-garden",
4
- "sourceCommit": "363dd9b54506b6cc67415e2bf76a856b743b7534",
4
+ "sourceCommit": "b99053b872172912413d4695863db9396cade500",
5
5
  "common": {
6
6
  "codexSkills": [
7
7
  "craft-rpa",
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "flower-trellis",
3
- "version": "0.4.6",
3
+ "version": "0.4.7",
4
4
  "description": "一键安装/升级 Trellis 并自动融合 skill-garden 强化包(默认 Claude + agents)",
5
5
  "type": "module",
6
6
  "bin": {
@@ -58,9 +58,9 @@
58
58
  "commit-and-tag-version": "^12.7.3"
59
59
  },
60
60
  "flowerReleaseNotes": {
61
- "version": "0.4.6",
61
+ "version": "0.4.7",
62
62
  "source": "CHANGELOG.md",
63
- "body": "### 新功能 Features\n\n* **update:** 注入跨版本更新摘要并联动 push 确认 ([bbdb23c](https://github.com/SilentFlower/flower-trellis/commit/bbdb23c33eac9948235c65f6be6f18764f90a88d))",
63
+ "body": "### 🐛 修复 Bug Fixes\n\n* **flower:** 同步 brainstorm planning gate ([3d2944e](https://github.com/SilentFlower/flower-trellis/commit/3d2944ebec9ae040a064d9974b170f046f4bf845))\n* **flower:** 同步 workflow 提示压缩 ([b7083ce](https://github.com/SilentFlower/flower-trellis/commit/b7083ce553ed4fda8190e30d7882358f70cb0449))",
64
64
  "truncated": false
65
65
  }
66
66
  }
@@ -31,7 +31,8 @@ function stripSkillOverride(value, name) {
31
31
  }
32
32
 
33
33
  /**
34
- * 把 override 块插入到 frontmatter 后、正文前;无 frontmatter 时插到文件顶部。
34
+ * 把 override 块插入到 frontmatter 后、正文前;无 frontmatter 时优先插到首个 H1 后,
35
+ * 都没有才插到文件顶部。
35
36
  *
36
37
  * @param {string} value 已清理旧块的目标文本
37
38
  * @param {string} block override 块
@@ -44,6 +45,13 @@ function injectAfterFrontmatter(value, block) {
44
45
  const after = value.slice(m[0].length).replace(/^\n+/, "");
45
46
  return value.slice(0, m[0].length) + "\n" + b + "\n\n" + after;
46
47
  }
48
+ // 无 frontmatter 的 command 文件首行标题常被平台当作命令描述来源;
49
+ // override 放到标题后,避免把高优先级 override 标题暴露成命令名。
50
+ const h1 = /^# [^\n]*\n/.exec(value);
51
+ if (h1) {
52
+ const after = value.slice(h1[0].length).replace(/^\n+/, "");
53
+ return value.slice(0, h1[0].length) + "\n" + b + "\n\n" + after;
54
+ }
47
55
  return b + "\n\n" + value.replace(/^\n+/, "");
48
56
  }
49
57