geneseed 5.5.0 → 5.5.1

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.
Files changed (69) hide show
  1. package/CHANGELOG.md +17 -0
  2. package/README.md +9 -9
  3. package/SHIPPED.md +1 -1
  4. package/docs/extending.md +1 -1
  5. package/docs/guides/install.md +5 -5
  6. package/docs/reference/profiles.md +84 -0
  7. package/harness.config.json +1 -1
  8. package/js/build/export.mjs +11 -16
  9. package/js/hosts/hooks.mjs +4 -1
  10. package/package.json +1 -1
  11. package/profiles/autonomous.json +14 -0
  12. package/profiles/corporate.json +14 -0
  13. package/profiles/mentor.json +13 -0
  14. package/profiles/minimal.json +15 -0
  15. package/src/agents/devops-engineer.md +6 -5
  16. package/src/agents/kubernetes-specialist.md +3 -1
  17. package/src/agents/spring-boot-engineer.md +3 -3
  18. package/src/agents/spring-reviewer.md +4 -2
  19. package/src/skills/codebase-design.md +2 -2
  20. package/src/skills/diagram-design/SKILL.md +12 -7
  21. package/src/skills/diagram-design/VENDOR.md +19 -14
  22. package/src/skills/diagram-design/references/layout-budget.md +7 -1
  23. package/src/skills/diagram-design/references/onboarding.md +1 -1
  24. package/src/skills/diagram-design/references/primitives-core.md +1 -1
  25. package/src/skills/diagram-design/references/semantic-patterns.md +1 -1
  26. package/src/skills/diagram-design/references/style-guide.md +1 -1
  27. package/src/skills/diagram-design/references/type-dependency.md +69 -0
  28. package/src/skills/diagram-design/references/type-event-modeling.md +36 -0
  29. package/src/skills/diagram-design/references/type-requirement-traceability.md +45 -0
  30. package/src/skills/diagram-design/references/type-sbgn.md +80 -0
  31. package/src/skills/diagram-design/references/type-scatter.md +1 -1
  32. package/src/skills/diagram-design/references/type-steps.md +180 -0
  33. package/src/skills/diagram-design/references/type-unit-grid.md +63 -0
  34. package/src/skills/geneseed-code-review.md +2 -1
  35. package/src/skills/parallel-agents.md +1 -1
  36. package/src/skills/react-view-transitions/AGENTS.md +26 -26
  37. package/src/skills/react-view-transitions/SKILL.md +1 -1
  38. package/src/skills/react-view-transitions/VENDOR.md +12 -3
  39. package/src/skills/spring-boot/cloud.md +19 -1
  40. package/src/skills/spring-boot/logging.md +5 -3
  41. package/src/skills/spring-boot/web.md +27 -2
  42. package/src/skills/spring-boot.md +4 -2
  43. package/src/skills/spring-data-jpa.md +18 -4
  44. package/src/skills/spring-testing.md +4 -0
  45. package/src/upstreams.json +4 -4
  46. package/web/dist/assets/{Activity-LKsDzsnK.js → Activity-bUSPM3Dl.js} +1 -1
  47. package/web/dist/assets/{ActivityDetail-DhwTfPm6.js → ActivityDetail-AHRoXeFN.js} +1 -1
  48. package/web/dist/assets/{Bonsai-CePMLrpr.js → Bonsai-B5XY9pid.js} +1 -1
  49. package/web/dist/assets/{CopyButton-CuoaJ6k7.js → CopyButton-C0wwRR9H.js} +1 -1
  50. package/web/dist/assets/{Dashboard-D1VKuvLe.js → Dashboard-DPAA96EI.js} +2 -2
  51. package/web/dist/assets/{Docs-CUrHJKaU.js → Docs-CpGDq165.js} +1 -1
  52. package/web/dist/assets/{Factories-Ci1XfmSy.js → Factories-cNhETa5Q.js} +1 -1
  53. package/web/dist/assets/{GreenhouseView-CNfH2lPK.js → GreenhouseView-DO3ADmT9.js} +1 -1
  54. package/web/dist/assets/{Installs-ZHQZM1iS.js → Installs-D3dDyq10.js} +1 -1
  55. package/web/dist/assets/{JournalView-CMsWahkS.js → JournalView--W-Lw_V9.js} +1 -1
  56. package/web/dist/assets/{Kits-CkliyK0n.js → Kits-ChfehxY6.js} +1 -1
  57. package/web/dist/assets/Laws-CNxYmV36.js +4 -0
  58. package/web/dist/assets/{LibRows-Brv0Hmkr.js → LibRows-CiPvsRYB.js} +1 -1
  59. package/web/dist/assets/{Library-BYmiLjJr.js → Library-BxQvbwVU.js} +1 -1
  60. package/web/dist/assets/{Loops-ytm-xJ68.js → Loops-DAblKkyq.js} +1 -1
  61. package/web/dist/assets/{OperatorHudView-KaqCx_HP.js → OperatorHudView-QtSaSyYm.js} +1 -1
  62. package/web/dist/assets/{Personal-CmcnRKYE.js → Personal-DxSGv638.js} +1 -1
  63. package/web/dist/assets/{Settings-tUrpdOfq.js → Settings-CgTbhvRv.js} +1 -1
  64. package/web/dist/assets/{Tabs-CD7etcGD.js → Tabs-Cf4X1hdz.js} +1 -1
  65. package/web/dist/assets/{index-COKbhFNo.js → index-B_pTwZra.js} +3 -3
  66. package/web/dist/assets/index-CzfImiYb.css +1 -0
  67. package/web/dist/index.html +2 -2
  68. package/web/dist/assets/Laws-hh4O3HZ4.js +0 -4
  69. package/web/dist/assets/index-Ba_Kg3Af.css +0 -1
package/CHANGELOG.md CHANGED
@@ -8,6 +8,23 @@ label. For the capability ↔ spec map, see [SHIPPED.md](SHIPPED.md).
8
8
 
9
9
  ## [Unreleased]
10
10
 
11
+ ## [5.5.1] — 2026-11-17
12
+
13
+ ### Added
14
+
15
+ - **Four more Shipped install profiles.** `mentor` (a learner: mentor posture, prudent trust, every gate armed, the ownership gate included), `corporate` (a locked-down work PC: prudent trust, `--strict-consent`, no login item), `autonomous` (a fluent user who delegates: expert posture, foreman mode, aggressive trust, 1m context, the deletion and secrets gates armed) and `minimal` (a small model or a tight token budget: assistant posture, the `craft` and `rigor` packs only, eight niche skills left out). [Install profiles](docs/reference/profiles.md) gives the reason for each choice.
16
+
17
+ ### Changed
18
+
19
+ - **The Constitution page's doctrine switches apply on their own.** There is no Apply button to click: once the switches have been still for a moment, the install rebuilds once with the whole selection, so a burst of clicks or a pack switch is still one rebuild. A plain rule rebuilds without a dialog; a change that also removes a gate or a hook, or offers to re-arm one, still asks, and a no puts the switches back. The switches lock while their rebuild runs. A rebuild the server refuses keeps the change on screen with Retry and Revert, and is never retried on its own.
20
+ - `geneseed export` writes the harness release as the version of every entity, skills and agents included; it no longer reads the never-bumped per-entity versions of `registry.json`.
21
+ - `geneseed export` writes the constitution as a side package: a skill or an agent that cites a law no longer lists the `laws` entity in its `requires`, so a library can take the skills and agents without the laws. The `laws` entity itself is still exported (the profile's `laws: false` still leaves it out).
22
+
23
+ ### Fixed
24
+
25
+ - The one-line `starter` install failed on a second run, and Windows PowerShell 5.1 could not run it: `git clone` refuses an existing folder, and PowerShell 5.1 has no `&&`. The README and the install guide now give a line that clones the first time and pulls on later runs, so the same line installs, updates and repairs. On Windows the line runs through `cmd /c`, so it works in cmd, PowerShell 5.1 and PowerShell 7.
26
+ - `geneseed export` lost the constitution's `LAWS.md` index on Windows and macOS: it shared a folder with `laws.md`, the same file on a case-insensitive file system. The laws file is now `laws-<stem>.md` (`laws-universal.md`), and a test refuses any two exported paths that differ only by case. A library that took an earlier export holds a stale `laws.md`: remove it by hand.
27
+
11
28
  ## [5.5.0] — 2026-11-15
12
29
 
13
30
  **Upgrade note.** The new behaviour reaches an install only when it is emitted again: `geneseed upgrade` (or `rebuild-all`) does it for every active install. An existing install reads as `--context-window 200k`, so on a machine whose model has a 1M-token window, rebuild it once with `--context-window 1m` or apply a profile that says `1m`. Project notes written by earlier versions move to `<project>/notebook/` on that rebuild, with no loss.
package/README.md CHANGED
@@ -30,14 +30,14 @@
30
30
  **New here? One line, no questions** — it clones Geneseed and installs the [`starter` profile](#-new-to-ai-agents-one-line-no-questions). You need git and Node ≥ 22.7.
31
31
 
32
32
  ```bat
33
- git clone https://github.com/Arylmera/Geneseed.git && cd Geneseed && .\install.cmd --profile starter
33
+ cmd /c "(if exist Geneseed (git -C Geneseed pull) else (git clone https://github.com/Arylmera/Geneseed.git)) && cd Geneseed && .\install.cmd --profile starter"
34
34
  ```
35
35
 
36
36
  ```bash
37
- git clone https://github.com/Arylmera/Geneseed.git && cd Geneseed && ./install --profile starter
37
+ { [ -d Geneseed ] && git -C Geneseed pull || git clone https://github.com/Arylmera/Geneseed.git; } && cd Geneseed && ./install --profile starter
38
38
  ```
39
39
 
40
- The first line is for Windows (cmd or PowerShell 7), the second for macOS and Linux. Windows PowerShell 5.1 has no `&&`: there, put `;` in its place. Want to choose the theme, the host and the rest yourself? Run the wizard instead:
40
+ The first line is for Windows (cmd, PowerShell 5.1 or 7), the second for macOS and Linux. Run it from the folder that gets the clone. It clones the first time and pulls on later runs, so the same line installs, updates and repairs. Want to choose the theme, the host and the rest yourself? Run the wizard instead:
41
41
 
42
42
  ```bash
43
43
  npx geneseed setup
@@ -88,21 +88,21 @@ geneseed-build --theme imperial # Warhammer 40k voice, identical structure
88
88
 
89
89
  ### 🧭 New to AI agents? One line, no questions
90
90
 
91
- Clone Geneseed where you like, then run one command at the root of the clone. It installs the **`starter`** profile into every AI coding tool it finds (OpenCode, IBM Bob, Claude Code, OpenClaude), puts `geneseed` on your `PATH`, sets the web console to start at login, starts it, runs the health check and prints the console's address. It asks nothing.
91
+ Run one line from the folder that gets the clone. It clones Geneseed the first time and pulls on later runs, so the same line installs, updates and repairs. It installs the **`starter`** profile into every AI coding tool it finds (OpenCode, IBM Bob, Claude Code, OpenClaude), puts `geneseed` on your `PATH`, sets the web console to start at login, starts it, runs the health check and prints the console's address. It asks nothing.
92
92
 
93
- Windows (cmd or PowerShell 7; on Windows PowerShell 5.1, put `;` in place of each `&&`):
93
+ Windows (cmd, PowerShell 5.1 or 7):
94
94
 
95
95
  ```bat
96
- git clone https://github.com/Arylmera/Geneseed.git && cd Geneseed && .\install.cmd --profile starter
96
+ cmd /c "(if exist Geneseed (git -C Geneseed pull) else (git clone https://github.com/Arylmera/Geneseed.git)) && cd Geneseed && .\install.cmd --profile starter"
97
97
  ```
98
98
 
99
99
  macOS, Linux:
100
100
 
101
101
  ```bash
102
- git clone https://github.com/Arylmera/Geneseed.git && cd Geneseed && ./install --profile starter
102
+ { [ -d Geneseed ] && git -C Geneseed pull || git clone https://github.com/Arylmera/Geneseed.git; } && cd Geneseed && ./install --profile starter
103
103
  ```
104
104
 
105
- `starter` is a neutral voice, a candid peer posture, the agent working each task itself, and every guardrail armed except the ownership gate. Run the same line again to update and repair. `geneseed profile save <name> --from claude-global` writes your own install as a Personal profile (in `~/.config/geneseed/profiles/`); with `--shipped`, from a clone, it writes [`profiles/`](profiles/) for a pull request, which is how a team adds one. `geneseed profile apply <name>` applies a profile globally, or with `--project <dir>` to one project folder. See [Install profiles](docs/reference/profiles.md).
105
+ `starter` is a neutral voice, a candid peer posture, the agent working each task itself, and every guardrail armed except the ownership gate. `geneseed profile save <name> --from claude-global` writes your own install as a Personal profile (in `~/.config/geneseed/profiles/`); with `--shipped`, from a clone, it writes [`profiles/`](profiles/) for a pull request, which is how a team adds one. `geneseed profile apply <name>` applies a profile globally, or with `--project <dir>` to one project folder. See [Install profiles](docs/reference/profiles.md).
106
106
 
107
107
  ### ⚡ The short way — `npx`
108
108
 
@@ -414,7 +414,7 @@ Some skills and agents are adapted from other open-source skill collections, und
414
414
  - `diagram-design` — [cathrynlavery/diagram-design](https://github.com/cathrynlavery/diagram-design) by Cathryn Lavery (MIT): the vendored `diagram-design` skill, trimmed (no Python helpers, two of its examples).
415
415
  - `mattpocock-skills` — [mattpocock/skills](https://github.com/mattpocock/skills) by Matt Pocock (MIT): the vocabulary and procedures behind codebase-design, wayfinder, prototype, develop, handoff and research, the spec pass of geneseed-code-review, the question rounds and questionnaire of brainstorm, the slicing of plan, the reproduction step of debug, the integration-branch rules of parallel-agents, the trigger-pruning of skill-forge, and the dated-history rule of wiki.
416
416
  - `mattpocock-skills-frozen` — [mattpocock/skills](https://github.com/mattpocock/skills) by Matt Pocock (MIT), pinned at 391a2701, the last commit carrying `resolving-merge-conflicts` (deleted upstream since): the merge-conflict step of git-history, from a skill upstream has since deleted, so the weekly drift check skips it.
417
- - `mattpocock-skills-2509` — [mattpocock/skills](https://github.com/mattpocock/skills) by Matt Pocock (MIT), a second key for the same repo, at the same pin as `mattpocock-skills` today (5beb35b0): the PR body of ship and the guardrail clause of rule.
417
+ - `mattpocock-skills-2509` — [mattpocock/skills](https://github.com/mattpocock/skills) by Matt Pocock (MIT), a second key for the same repo, pinned at 5beb35b0: the PR body of ship and the guardrail clause of rule.
418
418
  - `superpowers` — [obra/superpowers](https://github.com/obra/superpowers) by Jesse Vincent (MIT): the core of debug, the size step and builder check of brainstorm, the plan rules of plan, the reviewer brief behind geneseed-code-review and the reviewer agent, the review-handling steps of review-response, and the ruling format of pipeline and parallel-agents.
419
419
  - `vibe-wise` — [nykooi1/vibe-wise](https://github.com/nykooi1/vibe-wise) by Noah Kim (MIT): the state file, levels and checkpoints of learn-mode, and the checkpoint rules ownership shares.
420
420
  - `piomin-spring-boot` — [piomin/claude-ai-spring-boot](https://github.com/piomin/claude-ai-spring-boot) by Piotr Mińkowski (Apache-2.0): the Spring Boot kit's skills and agents, rewritten and condensed for Boot 4.
package/SHIPPED.md CHANGED
@@ -20,7 +20,7 @@ historical references, not files distributed with the repo.
20
20
  > **On version numbers.** [`harness.config.json`](harness.config.json)'s `version` is a
21
21
  > human-readable label only. The canonical identity of an installed harness is the
22
22
  > **source fingerprint** stamped in `.geneseed-version` (see `geneseed version`), not the
23
- > string. Current label: **5.5.0**.
23
+ > string. Current label: **5.5.1**.
24
24
 
25
25
  ## Capabilities
26
26
 
package/docs/extending.md CHANGED
@@ -124,7 +124,7 @@ Notes that matter:
124
124
  held to it as a sum) because `node --test` over a glob that matches nothing prints `# tests 0`
125
125
  and **exits 0**. Node's default reporter changed between majors, which is how a green 1043-test
126
126
  suite once reported zero.
127
- - `--cli` runs the ~320-cell CLI/hook matrix against the expectations written in each cell —
127
+ - `--cli` runs the ~280-cell CLI/hook matrix against the expectations written in each cell —
128
128
  the only end-to-end gate on what a verb prints. CI runs it in `node-cells` on both platforms.
129
129
  A cell that needs a broken checkout plants its fault as an **edit of the real file**
130
130
  (`{ "edit": [[find, replace]] }` in `tests/helpers/matrix/cli.*.json`), never as a copy of it;
@@ -10,21 +10,21 @@ Geneseed builds a harness (the rules, agents, skills and hooks your AI coding to
10
10
 
11
11
  ## New to AI agents: one line, no questions
12
12
 
13
- You need **git** and **Node 22.7 or newer**. Clone Geneseed where you like, then run the installer with the `starter` profile:
13
+ You need **git** and **Node 22.7 or newer**. Run one line from the folder that gets the clone. It clones Geneseed the first time and pulls on later runs, so the same line installs, updates and repairs. It runs the installer with the `starter` profile:
14
14
 
15
- Windows (cmd or PowerShell 7; on Windows PowerShell 5.1, put `;` in place of each `&&`):
15
+ Windows (cmd, PowerShell 5.1 or 7):
16
16
 
17
17
  ```bat
18
- git clone https://github.com/Arylmera/Geneseed.git && cd Geneseed && .\install.cmd --profile starter
18
+ cmd /c "(if exist Geneseed (git -C Geneseed pull) else (git clone https://github.com/Arylmera/Geneseed.git)) && cd Geneseed && .\install.cmd --profile starter"
19
19
  ```
20
20
 
21
21
  macOS, Linux:
22
22
 
23
23
  ```bash
24
- git clone https://github.com/Arylmera/Geneseed.git && cd Geneseed && ./install --profile starter
24
+ { [ -d Geneseed ] && git -C Geneseed pull || git clone https://github.com/Arylmera/Geneseed.git; } && cd Geneseed && ./install --profile starter
25
25
  ```
26
26
 
27
- It asks nothing. It installs the harness globally into every AI coding tool it finds (OpenCode, IBM Bob, Claude Code, OpenClaude), puts `geneseed` on your `PATH`, sets the web console to start at login, starts it, runs the health check, and prints a summary with the console's address. With no tool on the machine it installs nothing and prints where to get each one. Run the same line again to update and repair. `starter` is a neutral voice, a candid peer posture, and every guardrail armed except the ownership gate. [Install profiles](../reference/profiles.md) has the details: how to save your own setup as a Personal profile, apply a profile to one project with `geneseed profile apply`, and how a team adds a Shipped one.
27
+ It asks nothing. It installs the harness globally into every AI coding tool it finds (OpenCode, IBM Bob, Claude Code, OpenClaude), puts `geneseed` on your `PATH`, sets the web console to start at login, starts it, runs the health check, and prints a summary with the console's address. With no tool on the machine it installs nothing and prints where to get each one. `starter` is a neutral voice, a candid peer posture, and every guardrail armed except the ownership gate. [Install profiles](../reference/profiles.md) has the details: how to save your own setup as a Personal profile, apply a profile to one project with `geneseed profile apply`, and how a team adds a Shipped one.
28
28
 
29
29
  ## The 5-minute path
30
30
 
@@ -116,6 +116,90 @@ A gate that is off in the global install runs nowhere, so the line does not name
116
116
 
117
117
  `peer` and not `expert`, because `expert` does not explain the basics a new user needs; not `mentor`, because it checks understanding with questions. Every gate stays armed except the ownership gate, whose first-write stop is heavy for a new user. `200k`, because most machines run a 200k-token model, a locked-down work PC included; `guillaume-claude-global` sets `1m` for a 1M-token model.
118
118
 
119
+ ## `mentor`
120
+
121
+ ```json
122
+ {
123
+ "description": "For a learner (a junior, a student): neutral voice, mentor posture, every guardrail, the ownership gate included.",
124
+ "hosts": ["opencode", "bob", "claude", "openclaude"],
125
+ "theme": "neutral",
126
+ "footprint": "lean",
127
+ "posture": "mentor",
128
+ "mode": "direct",
129
+ "trust": "prudent",
130
+ "doctrines": "all",
131
+ "kits": [],
132
+ "contextWindow": "200k",
133
+ "login": true
134
+ }
135
+ ```
136
+
137
+ `mentor`, because it explains the reasoning behind each step and checks understanding before it moves on. A learner needs that; a fluent user does not. `prudent`, so that a loop stops for a person at a lower risk score ([Loop risk](../concepts/loop-risk.md)). The profile turns no gate off. The ownership gate stays armed: in an application repo, it asks the learner to set up ownership, and then the learner decides each part of the code ([Ownership](../concepts/ownership.md)).
138
+
139
+ ## `corporate`
140
+
141
+ ```json
142
+ {
143
+ "description": "For a locked-down work PC: neutral voice, every guardrail except the ownership gate, strict consent, no login item.",
144
+ "hosts": ["opencode", "bob", "claude", "openclaude"],
145
+ "theme": "neutral",
146
+ "footprint": "lean",
147
+ "posture": "peer",
148
+ "mode": "direct",
149
+ "trust": "prudent",
150
+ "doctrines": "all",
151
+ "kits": [],
152
+ "contextWindow": "200k",
153
+ "gates": { "ownershipGate": false, "ownBranchPass": false },
154
+ "login": false
155
+ }
156
+ ```
157
+
158
+ `prudent`, because a work machine holds code that other people answer for. `ownBranchPass: false` is `--strict-consent`: every commit and push asks, on your own branch too, not only on the main branch. The ownership gate is off, as in `starter`: a team that wants ownership writes `mode: on` in its repo's `AGENTS.md`, and that section binds with the gate off. `login: false`, because a policy often blocks a login item on a managed PC. Start the console by hand with `geneseed web start`.
159
+
160
+ ## `autonomous`
161
+
162
+ ```json
163
+ {
164
+ "description": "For a fluent user who delegates: expert posture, foreman mode, aggressive loops, every guardrail except the ownership gate.",
165
+ "hosts": ["opencode", "bob", "claude", "openclaude"],
166
+ "theme": "neutral",
167
+ "footprint": "lean",
168
+ "posture": "expert",
169
+ "mode": "foreman",
170
+ "trust": "aggressive",
171
+ "doctrines": "all",
172
+ "kits": [],
173
+ "contextWindow": "1m",
174
+ "gates": { "ownershipGate": false },
175
+ "login": true
176
+ }
177
+ ```
178
+
179
+ `expert` and `foreman`, because this user gives a task and lets background workers do it. `aggressive` lets a loop accept a higher risk score before it stops for a person ([Loop risk](../concepts/loop-risk.md)). The deletion gate and the secrets gate stay armed: a worker that runs alone must not lose work or leak a key. Only the ownership gate is off, because its first-write stop interrupts the work. `1m`, because a user who delegates large tasks usually runs a 1M-token model. On a 200k-token model, save your own profile with `contextWindow` set to `200k`.
180
+
181
+ ## `minimal`
182
+
183
+ ```json
184
+ {
185
+ "description": "For a small model or a tight token budget: the smallest always-on context, two doctrine packs, niche skills left out.",
186
+ "hosts": ["opencode", "bob", "claude", "openclaude"],
187
+ "theme": "neutral",
188
+ "footprint": "lean",
189
+ "posture": "assistant",
190
+ "mode": "direct",
191
+ "trust": "balanced",
192
+ "doctrines": ["craft", "rigor"],
193
+ "excludeSkills": ["brick-forge", "bruno", "daydream", "factory", "factory-forge", "forge-mcp", "prototype", "react-view-transitions"],
194
+ "kits": [],
195
+ "contextWindow": "200k",
196
+ "gates": { "ownershipGate": false },
197
+ "login": true
198
+ }
199
+ ```
200
+
201
+ Each choice makes the always-on context smaller. `assistant` does what you ask and stops, so a small model gets a narrow task and few chances to drift. `craft` and `rigor` are the packs about how code is written and proven; the other packs stay in `doctrines/` for reference. The excluded skills serve rare tasks. A skill that the harness links to cannot be excluded, so `council`, `deck`, `pipeline` and `workflow` stay. Every gate except the ownership gate stays armed, because a gate costs no context. To measure the gain on your machine, build `starter` and `minimal` into two empty folders with `geneseed profile apply <name> --project <dir> --hosts claude`, then compare the size of each `CLAUDE.md` and the number of skills.
202
+
119
203
  ## Save your own
120
204
 
121
205
  ```bash
@@ -2,5 +2,5 @@
2
2
  "_comment": "version is a human-readable release label only; the canonical identity of an installed harness is the source fingerprint stamped in .geneseed-version (see `harness version`), not this string.",
3
3
  "theme": "neutral",
4
4
  "product": "Geneseed",
5
- "version": "5.5.0"
5
+ "version": "5.5.1"
6
6
  }
@@ -31,8 +31,9 @@
31
31
  * An agent keeps the whole of its seven sections, step 0 and Self-improvement included: that
32
32
  * shape is the contract a persona library reads;
33
33
  * - host keys (`disable-model-invocation`, `disallowedTools`, ...) are not written at all.
34
- * A citation of a law or doctrine stays the principle's name, and the entity `requires` the
35
- * exported constitution, which is where that name is defined.
34
+ * A citation of a law or doctrine stays the principle's name, and the entity does NOT `require`
35
+ * the exported constitution: the laws are a side package a library may take or leave, and a
36
+ * skill or an agent installs without them (the user decided, 2026-11).
36
37
  *
37
38
  * NOT EXPORTED: an alias (a second name for one skill), the authoring templates, and a vendored
38
39
  * folder skill. The last is a third party's folder shipped verbatim under its own licence and
@@ -88,7 +89,6 @@ const PROFILE_DEFAULTS = {
88
89
  };
89
90
  const LIST_KEYS = ['compatibleTools', 'exclude'];
90
91
  const NAME_RE = /^[a-z0-9]+(-[a-z0-9]+)*$/;
91
- const CITES_RE = /\{\{(?:LAW|DOCTRINE):[a-z0-9-]+\}\}/;
92
92
 
93
93
  /** A refusal: the message is the whole of what the user needs, so no stack is printed. */
94
94
  class ExportError extends Error {}
@@ -161,13 +161,6 @@ function metadataYaml(m) {
161
161
 
162
162
  const frontmatter = (name, description) => `---\nname: ${name}\ndescription: ${yamlStr(description)}\n---\n\n`;
163
163
 
164
- /** The registry's row for one entity: its version and nothing else is read. */
165
- function registryVersions() {
166
- try {
167
- return JSON.parse(readText(path.join(ROOT, 'registry.json'))).entities ?? {};
168
- } catch { return {}; }
169
- }
170
-
171
164
  /**
172
165
  * Rewrite every reference to another exported entity to its exported name, and collect them.
173
166
  * Two spellings carry a reference in the sources: a Markdown link to the entity's spec
@@ -226,7 +219,6 @@ export function planExport({ profile = { ...PROFILE_DEFAULTS }, theme = 'neutral
226
219
  // a lean line.
227
220
  const { items } = renderAll(cfg, theme, { footprint: 'full' });
228
221
  const srcRel = (it) => relPosix(cfg.src, it.src);
229
- const versions = registryVersions();
230
222
  const release = sourceReleaseVersion(cfg);
231
223
 
232
224
  // The roster: flat specs under skills/ and agents/, minus templates and what the profile drops.
@@ -289,11 +281,12 @@ export function planExport({ profile = { ...PROFILE_DEFAULTS }, theme = 'neutral
289
281
  description = agentDescription(s.it.text);
290
282
  body = link(body);
291
283
  }
292
- if (profile.laws && CITES_RE.test(readText(s.it.src))) requires.add(lawsName);
293
284
  files.push({ rel: `${dir}/${s.kind === 'skill' ? 'SKILL.md' : 'AGENT.md'}`, text: frontmatter(s.name, description) + body });
294
285
  files.push({ rel: `${dir}/metadata.yaml`, text: metadataYaml({
295
286
  ...base, type: s.kind, requires: [...requires].sort(),
296
- version: versions[`${s.kind === 'skill' ? 'skills' : 'agents'}/${s.stem}`]?.version ?? release,
287
+ // Every entity carries the harness release, as the laws do: the set ships as one, and the
288
+ // per-entity versions in registry.json were never bumped (the user decided, 2026-11).
289
+ version: release,
297
290
  ...creditsOf(s.it.src, cfg.src),
298
291
  }) });
299
292
  entities.push({ type: s.kind, name: s.name, source: srcRel(s.it), path: dir });
@@ -325,15 +318,17 @@ export function planExport({ profile = { ...PROFILE_DEFAULTS }, theme = 'neutral
325
318
  : tier === 'laws'
326
319
  ? ['Laws', `the invariants, always in force (${rules} rules)`]
327
320
  : ['Ontology', 'how the agent thinks, decides and behaves'];
328
- const file = tier === 'doctrines' ? `doctrines-${stem}.md` : `${tier}.md`;
321
+ // `laws-<stem>.md`, never `laws.md`: beside the `LAWS.md` index that name is the SAME file
322
+ // on a case-insensitive file system (Windows, macOS), and the index was silently lost.
323
+ const file = tier === 'ontology' ? 'ontology.md' : `${tier}-${stem}.md`;
329
324
  parts.push({ rank: ['ontology', 'laws'].indexOf(tier) >= 0 ? ['ontology', 'laws'].indexOf(tier) : 2 + packs.indexOf(stem), file, title, what });
330
325
  files.push({ rel: `${dir}/${file}`, text: `# ${title}\n\n${body}\n` });
331
326
  }
332
327
  parts.sort((x, y) => x.rank - y.rank);
333
328
  files.push({ rel: `${dir}/LAWS.md`, text: [
334
329
  `# ${lawsName}`, '',
335
- 'The constitution the skills and agents of this collection cite by principle name. One '
336
- + 'file per category:', '',
330
+ 'The constitution the skills and agents of this collection cite by principle name. It is a '
331
+ + 'side package: no skill or agent requires it. One file per category:', '',
337
332
  ...parts.map((x) => `- \`${x.file}\` — ${x.title}: ${x.what}`), '',
338
333
  ].join('\n') });
339
334
  files.push({ rel: `${dir}/metadata.yaml`, text: metadataYaml({
@@ -1583,8 +1583,11 @@ export const cmdRuleGate = guardGate(ruleGate, 'rule-gate');
1583
1583
  * for PreToolUse), so `gitDecide`/`ruleDecide` keep reading the Claude shape unchanged and
1584
1584
  * never learn Bob exists. A no-op for an already-Claude-shaped payload: every field it sets
1585
1585
  * is only filled in when the Claude-named field is absent.
1586
+ *
1587
+ * Exported for the in-process loss-table rows of tests/unit/hook_gates.test.mjs, which hand
1588
+ * Bob's own payload shape to `gitDecide` without a spawn per row.
1586
1589
  */
1587
- function normaliseBobPayload(payload) {
1590
+ export function normaliseBobPayload(payload) {
1588
1591
  if (!payload || typeof payload !== 'object') return payload;
1589
1592
  const out = { ...payload };
1590
1593
  if (!out.tool_input && payload.input && typeof payload.input === 'object') {
package/package.json CHANGED
@@ -1,7 +1,7 @@
1
1
  {
2
2
  "//": "Every field here is argued and gated in tests/unit/package_manifest.test.mjs. `files`, `bin`, `exports` and `engines` each have a test that fails when reality drifts from this file; `version` mirrors harness.config.json and is gated too. Read that docblock before editing this one.",
3
3
  "name": "geneseed",
4
- "version": "5.5.0",
4
+ "version": "5.5.1",
5
5
  "description": "A portable, theme-able harness you implant once and use everywhere to grow a disciplined AI coding agent.",
6
6
  "keywords": [
7
7
  "ai",
@@ -0,0 +1,14 @@
1
+ {
2
+ "description": "For a fluent user who delegates: expert posture, foreman mode, aggressive loops, every guardrail except the ownership gate.",
3
+ "hosts": ["opencode", "bob", "claude", "openclaude"],
4
+ "theme": "neutral",
5
+ "footprint": "lean",
6
+ "posture": "expert",
7
+ "mode": "foreman",
8
+ "trust": "aggressive",
9
+ "doctrines": "all",
10
+ "kits": [],
11
+ "contextWindow": "1m",
12
+ "gates": { "ownershipGate": false },
13
+ "login": true
14
+ }
@@ -0,0 +1,14 @@
1
+ {
2
+ "description": "For a locked-down work PC: neutral voice, every guardrail except the ownership gate, strict consent, no login item.",
3
+ "hosts": ["opencode", "bob", "claude", "openclaude"],
4
+ "theme": "neutral",
5
+ "footprint": "lean",
6
+ "posture": "peer",
7
+ "mode": "direct",
8
+ "trust": "prudent",
9
+ "doctrines": "all",
10
+ "kits": [],
11
+ "contextWindow": "200k",
12
+ "gates": { "ownershipGate": false, "ownBranchPass": false },
13
+ "login": false
14
+ }
@@ -0,0 +1,13 @@
1
+ {
2
+ "description": "For a learner (a junior, a student): neutral voice, mentor posture, every guardrail, the ownership gate included.",
3
+ "hosts": ["opencode", "bob", "claude", "openclaude"],
4
+ "theme": "neutral",
5
+ "footprint": "lean",
6
+ "posture": "mentor",
7
+ "mode": "direct",
8
+ "trust": "prudent",
9
+ "doctrines": "all",
10
+ "kits": [],
11
+ "contextWindow": "200k",
12
+ "login": true
13
+ }
@@ -0,0 +1,15 @@
1
+ {
2
+ "description": "For a small model or a tight token budget: the smallest always-on context, two doctrine packs, niche skills left out.",
3
+ "hosts": ["opencode", "bob", "claude", "openclaude"],
4
+ "theme": "neutral",
5
+ "footprint": "lean",
6
+ "posture": "assistant",
7
+ "mode": "direct",
8
+ "trust": "balanced",
9
+ "doctrines": ["craft", "rigor"],
10
+ "excludeSkills": ["brick-forge", "bruno", "daydream", "factory", "factory-forge", "forge-mcp", "prototype", "react-view-transitions"],
11
+ "kits": [],
12
+ "contextWindow": "200k",
13
+ "gates": { "ownershipGate": false },
14
+ "login": true
15
+ }
@@ -25,11 +25,12 @@ Builds CI/CD pipelines and delivery automation for Spring Boot applications.
25
25
  1. Detect the CI the project already uses (GitHub Actions, GitLab CI, Jenkins, other) and extend it; do not switch providers. With none present, ask the user which to use.
26
26
  2. Pipeline order: checkout, restore build cache, `./mvnw verify` (or `./gradlew build`), publish test results, build image, scan, push, deploy.
27
27
  3. Cache the Maven or Gradle repository keyed on the build files.
28
- 4. Gates: tests, a coverage threshold if the project has one, dependency and image vulnerability scan.
29
- 5. Build images with the project's chosen method (`spring-boot:build-image` by default); tag with version or commit SHA, never `latest` in production.
30
- 6. Stages dev, staging, prod, with approval before prod.
31
- 7. Secrets come from the CI secret store or a vault, never from files in the repository.
32
- 8. Lint the pipeline file before handing off.
28
+ 4. Gates: tests, a coverage threshold if the project has one, dependency and image vulnerability scan, `kustomize build` of every overlay when the project has them. Testcontainers needs a Docker daemon local to the job (on CircleCI the machine executor: a remote Docker's container ports are unreachable); confirm the integration tests ran (spring-testing skill).
29
+ 5. Docker Compose for local dependencies: follow the spring-boot skill's cloud side file (healthchecks).
30
+ 6. Build images with the project's chosen method (`spring-boot:build-image` by default); tag with version or commit SHA, never `latest` in production.
31
+ 7. Stages dev, staging, prod, with approval before prod.
32
+ 8. Secrets come from the CI secret store or a vault, never from files in the repository.
33
+ 9. Lint the pipeline file before handing off.
33
34
 
34
35
  ## Output contract
35
36
  - Files changed, the stage list and what was validated.
@@ -29,7 +29,9 @@ Deploys Spring Boot 4 applications on Kubernetes.
29
29
  5. Config: non-sensitive settings in a ConfigMap, credentials in a Secret; env vars bind through relaxed binding (`SPRING_DATASOURCE_URL`).
30
30
  6. Scale: HPA on CPU or custom metrics, a PodDisruptionBudget when replicas exceed one.
31
31
  7. Pod hardening: `runAsNonRoot`, no privilege escalation, pinned image tags.
32
- 8. Validate with a dry run or linter before handing off.
32
+ 8. What committed manifests may name, probes for dependencies deployed beside the service, and manifests for several platforms: follow the spring-boot skill's cloud side file.
33
+ 9. Check the target before giving deploy steps (`kubectl config current-context`, `oc whoami --show-server`); on OpenShift expose through Routes, not port forwarding.
34
+ 10. Validate with a dry run or linter before handing off.
33
35
 
34
36
  ## Output contract
35
37
  - Files changed and a one-line summary; say what was validated and what was not.
@@ -24,12 +24,12 @@ Implements Spring Boot 4 / Java 21+ features, following the spring-boot, spring-
24
24
  ## Procedure
25
25
  0. Read first, each if it exists — your durable lessons from prior dispatches ({{DOCTRINE:persist-insight}}): `{{PATH_MEMORY}}/agents/<your-name>.md`, then the feedback notes `{{PATH_NOTES}}/agent-notes/<your-name>.md` and `{{PATH_NOTES}}/agent-notes/_general.md`, then this repo's `{{PATH_PROJECT_NOTES}}/agent-notes/<your-name>.md` and `_general.md` there, which win where they disagree.
26
26
  1. Follow the project's existing layout and build tool; match its group id, packages and CI. Lombok is fine when the project already uses it.
27
- 2. Constructor injection; records for request and response DTOs; `@Valid` on request bodies; one `@RestControllerAdvice` for errors.
27
+ 2. Constructor injection; records for request and response DTOs; `@Valid` on request bodies; one `@RestControllerAdvice` for errors and `PagedModel` for lists, as the spring-boot skill's web side file shows.
28
28
  3. Services: `@Transactional(readOnly = true)` on the class, `@Transactional` on mutating methods. Never call a transactional or `@Async` method through `this`.
29
29
  4. Group related settings in `@ConfigurationProperties`, not scattered `@Value`.
30
30
  5. Outbound HTTP: `RestClient` for blocking calls, `@HttpExchange` interfaces for declarative clients; keep `WebClient` for reactive code only.
31
- 6. Persistence uses `jakarta.persistence`; never return entities from controllers.
32
- 7. Tests: positive and negative cases; `@MockitoBean` and `@MockitoSpyBean` for replaced beans; Testcontainers with `@ServiceConnection`.
31
+ 6. Persistence uses `jakarta.persistence`; never return entities from controllers. On a relational database the migration tool owns the schema (`ddl-auto: validate`): write the migration in the same change as the entity (spring-data-jpa skill).
32
+ 7. Tests: positive and negative cases; `@MockitoBean` and `@MockitoSpyBean` for replaced beans; Testcontainers with `@ServiceConnection`; what the integration tests must cover is in the spring-testing skill.
33
33
  8. Defaults worth setting: `server.shutdown=graceful` and `spring.lifecycle.timeout-per-shutdown-phase`.
34
34
  9. Run the affected build and tests before handing off ({{LAW:verify-before-asserting}}).
35
35
 
@@ -23,15 +23,17 @@ Spring-specific review only. Generic correctness, naming and structure belong to
23
23
 
24
24
  ## Procedure
25
25
  0. Read first, each if it exists — your durable lessons from prior dispatches ({{DOCTRINE:persist-insight}}): `{{PATH_MEMORY}}/agents/<your-name>.md`, then the feedback notes `{{PATH_NOTES}}/agent-notes/<your-name>.md` and `{{PATH_NOTES}}/agent-notes/_general.md`, then this repo's `{{PATH_PROJECT_NOTES}}/agent-notes/<your-name>.md` and `_general.md` there, which win where they disagree.
26
- 1. Read the changed files and the build file to learn the Boot version.
26
+ 1. Read the changed files and the build file to learn the Boot version. Review only the change: flag pre-existing code only where the change makes it worse, and never a choice the project documents on purpose.
27
27
  2. Check, in this order:
28
28
  - Injection: field `@Autowired`, circular dependencies.
29
29
  - Transactions: multi-step writes without `@Transactional`; self-invocation bypassing the proxy; transactions held across remote calls.
30
30
  - Persistence: N+1 on lazy collections (use `JOIN FETCH` or `@EntityGraph`), unbounded `findAll()`, entities returned from controllers, legacy `javax` persistence imports (the package is `jakarta.persistence`).
31
+ - Schema: an entity change with no migration in the same change, an edited applied migration, `ddl-auto: update` or `create` outside tests.
32
+ - Web: `Page`/`PageImpl` returned from a controller (use `PagedModel`) unless `spring.data.web.pageable.serialization-mode: via-dto` is set; a catch-all handler that returns the exception message; a stale update that does not answer 409.
31
33
  - Boot 4 API misuse: the removed Boot test-package mock and spy bean annotations (use `@MockitoBean`/`@MockitoSpyBean`), new `RestTemplate` code or blocking `WebClient` (use `RestClient`), new `TestRestTemplate` tests (prefer `RestTestClient`), `@DynamicPropertySource` where `@ServiceConnection` supports the container, `WebSecurityConfigurerAdapter`.
32
34
  - Test slices: `@SpringBootTest` where `@WebMvcTest` or `@DataJpaTest` would do; slices missing their collaborators.
33
35
  - Configuration: raw `@Value` for property groups, secrets in committed config.
34
- - Actuator: endpoints exposed beyond need, `show-details` open to anonymous callers.
36
+ - Actuator: endpoints exposed beyond need, `show-details` open to anonymous callers, probe paths blocked by security (cloud side file).
35
37
  3. Verify by running the tests rather than assuming ({{LAW:verify-before-asserting}}).
36
38
 
37
39
  ## Output contract
@@ -27,8 +27,8 @@ a design talk that keeps stumbling over ambiguous terms, or another {{SKILL}} ne
27
27
  3. Aim for **deep modules** — much behaviour behind a small interface: callers
28
28
  get leverage, maintainers get locality. To find what to deepen, start from
29
29
  `git log` hot spots — what keeps changing. Design a new or reshaped interface
30
- twice before choosing: 2–3 radically different shapes (in parallel subagents
31
- where the host runs them), compared on depth, locality and seam placement;
30
+ twice before choosing: 2–3 radically different shapes (in parallel fresh
31
+ subagents, never forks, where the host runs them), compared on depth, locality and seam placement;
32
32
  recommend one — a hybrid if parts combine — never a menu.
33
33
  4. Seams sparingly: the interface is the test surface. Internal seams for the
34
34
  module's own tests are fine but never exposed; a test that must reach past
@@ -1,6 +1,6 @@
1
1
  ---
2
2
  name: diagram-design
3
- description: "Branded HTML/SVG diagrams and charts — architecture, flowchart, sequence, ER, timeline, Gantt, Sankey and dozens more — with .drawio, Mermaid and .excalidraw import. Use when the user asks for a diagram, chart or visual map."
3
+ description: "Branded HTML/SVG diagrams and charts — architecture, flowchart, sequence, ER, timeline, Gantt, Sankey and dozens more — with .drawio, Mermaid, .excalidraw and PlantUML import. Use when the user asks for a diagram, chart or visual map."
4
4
  license: MIT
5
5
  metadata:
6
6
  version: "2.6"
@@ -10,7 +10,7 @@ metadata:
10
10
 
11
11
  Create diagrams as self-contained HTML files with inline SVG and an editorial design system.
12
12
 
13
- Forty-five visual types. Semantic patterns describe behavior; type references describe layout.
13
+ Fifty visual types. Semantic patterns describe behavior; type references describe layout.
14
14
 
15
15
  ---
16
16
 
@@ -49,7 +49,7 @@ Applied to schematics:
49
49
 
50
50
  ## 2. When to Use
51
51
 
52
- Use for any of the 45 visual types (§3) when a reader will learn more from a visual than from prose, a table, or a bulleted list.
52
+ Use for any of the 50 visual types (§3) when a reader will learn more from a visual than from prose, a table, or a bulleted list.
53
53
 
54
54
  **Don't use for:**
55
55
 
@@ -80,7 +80,7 @@ When behavior, state, enforcement, or risk carries the meaning, first load [`ref
80
80
 
81
81
  The pattern owns semantic primitives and its tighter budget; the type owns layout grammar. Use [`references/animation.md`](references/animation.md) only when motion is requested or materially clarifies ordered change; static remains the default.
82
82
 
83
- ### Visual-type guide (45)
83
+ ### Visual-type guide (50)
84
84
 
85
85
  | If you're showing… | Use | Reference |
86
86
  |---|---|---|
@@ -91,6 +91,7 @@ The pattern owns semantic primitives and its tighter budget; the type owns layou
91
91
  | Time-ordered messages between actors | **Sequence** | [type-sequence.md](references/type-sequence.md) |
92
92
  | States + transitions + guards | **State machine** | [type-state.md](references/type-state.md) |
93
93
  | Entities + fields + relationships | **ER / data model** | [type-er.md](references/type-er.md) |
94
+ | A short ordered method with numbered stages, icons, and brief explanations | **Steps** | [type-steps.md](references/type-steps.md) |
94
95
  | Events positioned in time | **Timeline** | [type-timeline.md](references/type-timeline.md) |
95
96
  | Cross-functional process with handoffs | **Swimlane** | [type-swimlane.md](references/type-swimlane.md) |
96
97
  | Two-axis positioning / prioritization | **Quadrant** | [type-quadrant.md](references/type-quadrant.md) |
@@ -125,10 +126,14 @@ The pattern owns semantic primitives and its tighter budget; the type owns layou
125
126
  | What a person does across stages of an experience, and how it feels | **User journey** | [type-journey.md](references/type-journey.md) |
126
127
  | What the customer does vs what the service hides, across stages, with a line of visibility | **Service blueprint** | [type-service-blueprint.md](references/type-service-blueprint.md) |
127
128
  | Where software runs — zones, hosts, artifacts, replicas, ports | **Deployment** | [type-deployment.md](references/type-deployment.md) |
128
- | What depends on what, with fan-in and cycles a tree cannot express | **Dependency graph** | [type-dependency.md](references/type-dependency.md) |
129
+ | What depends on what, with fan-in and cycles a tree cannot express; arc diagram (nodes on a line, height is skip-span) variant | **Dependency graph** | [type-dependency.md](references/type-dependency.md) |
129
130
  | Classes with operations, inheritance, composition (other UML routes elsewhere) | **UML class** | [type-uml-class.md](references/type-uml-class.md) |
130
131
  | Narrative backbone sliced into releases, with the cut line | **Story map** | [type-story-map.md](references/type-story-map.md) |
131
132
  | Physical tables: SQL types, constraints, indexes, column-level FKs | **Database schema** | [type-db-schema.md](references/type-db-schema.md) |
133
+ | A biological process as a bipartite species/process graph | **SBGN** | [type-sbgn.md](references/type-sbgn.md) |
134
+ | Source requirements, design elements, and tests linked by stable IDs, with gaps explicit | **Requirement traceability** | [type-requirement-traceability.md](references/type-requirement-traceability.md) |
135
+ | Commands, resulting events, and read models across ordered stages | **Event modeling** | [type-event-modeling.md](references/type-event-modeling.md) |
136
+ | One bounded whole counted in equal discrete units out of 100 | **Unit Grid** | [type-unit-grid.md](references/type-unit-grid.md) |
132
137
 
133
138
  Rules of thumb:
134
139
 
@@ -338,9 +343,9 @@ Every diagram ships in three variants (see `assets/`):
338
343
 
339
344
  ---
340
345
 
341
- ## 11. Importing an Existing Diagram (draw.io), Mermaid, and Excalidraw
346
+ ## 11. Importing an Existing Diagram (draw.io), Mermaid, Excalidraw, and PlantUML
342
347
 
343
- Sources: `.drawio*`, `.mmd`, `.mermaid`, Markdown containing a fenced `mermaid` block, and `.excalidraw`. Follow this section for "convert this", "redraw this diagram", and "make this presentable".
348
+ Sources: `.drawio*`, `.mmd`, `.mermaid`, Markdown containing a fenced `mermaid` block, `.excalidraw`, `.puml`, `.plantuml`, `.pu`, and Markdown containing a fenced `plantuml` / `puml` block. Follow this section for "convert this", "redraw this diagram", and "make this presentable".
344
349
 
345
350
  The short version:
346
351