geneseed 5.5.1 → 6.0.0

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 (88) hide show
  1. package/CHANGELOG.md +27 -0
  2. package/README.md +2 -2
  3. package/SHIPPED.md +2 -2
  4. package/adapters/opencode/README.md +7 -4
  5. package/docs/concepts/agents.md +8 -0
  6. package/docs/concepts/enforcement-table.md +1 -1
  7. package/docs/concepts/exclusions.md +1 -1
  8. package/docs/concepts/foreman-mode.md +5 -3
  9. package/docs/concepts/hooks.md +1 -1
  10. package/docs/concepts/kits.md +20 -2
  11. package/docs/concepts/loops.md +5 -0
  12. package/docs/concepts/skills.md +1 -1
  13. package/docs/concepts/superpowers-kit.md +79 -0
  14. package/docs/extending.md +11 -4
  15. package/docs/guides/install-scripted.md +1 -1
  16. package/docs/reference/cli.md +2 -0
  17. package/docs/reference/profiles.md +10 -5
  18. package/harness.config.json +1 -1
  19. package/js/README.md +2 -0
  20. package/js/build/args.mjs +4 -2
  21. package/js/build/driver.mjs +8 -8
  22. package/js/build/emit-claude.mjs +5 -3
  23. package/js/build/emit-opencode.mjs +2 -2
  24. package/js/build/generate.mjs +1 -1
  25. package/js/build/profile.mjs +15 -7
  26. package/js/build/render.mjs +81 -10
  27. package/js/build/source.mjs +75 -1
  28. package/js/cli-table.json +17 -2
  29. package/js/hosts/hooks.mjs +16 -5
  30. package/js/hosts/installs.mjs +50 -4
  31. package/js/hosts/native.mjs +104 -3
  32. package/js/hosts/opencode.mjs +13 -15
  33. package/js/inspect/checks-authoring.mjs +24 -1
  34. package/js/inspect/checks-build.mjs +43 -2
  35. package/js/inspect/checks-counts.mjs +72 -1
  36. package/js/inspect/diff.mjs +17 -3
  37. package/js/inspect/doctor.mjs +23 -2
  38. package/js/inspect/status.mjs +5 -1
  39. package/js/inspect/validate.mjs +1 -1
  40. package/js/loop/catalog.mjs +10 -0
  41. package/js/maintain/kit-migration.mjs +109 -0
  42. package/js/maintain/plugin-detect.mjs +150 -0
  43. package/js/maintain/profiles.mjs +46 -5
  44. package/js/maintain/setup.mjs +45 -11
  45. package/js/maintain/update.mjs +1 -1
  46. package/js/web/actions.mjs +28 -7
  47. package/js/web/api.mjs +3 -1
  48. package/js/web/profiles-api.mjs +11 -7
  49. package/package.json +1 -1
  50. package/profiles/autonomous.json +1 -0
  51. package/profiles/corporate.json +1 -1
  52. package/profiles/mentor.json +1 -0
  53. package/profiles/minimal.json +2 -1
  54. package/profiles/starter.json +1 -1
  55. package/src/AGENT.md.tmpl +1 -6
  56. package/src/agents/_template.md +4 -0
  57. package/src/agents/advocate.md +1 -0
  58. package/src/agents/architect.md +1 -0
  59. package/src/agents/developer.md +12 -4
  60. package/src/agents/empiricist.md +1 -0
  61. package/src/agents/explorer.md +1 -0
  62. package/src/agents/framer.md +1 -0
  63. package/src/agents/historian.md +1 -0
  64. package/src/agents/operator.md +1 -0
  65. package/src/agents/performance.md +2 -1
  66. package/src/agents/pragmatist.md +1 -0
  67. package/src/agents/react-reviewer.md +1 -0
  68. package/src/agents/researcher.md +1 -0
  69. package/src/agents/reviewer.md +1 -0
  70. package/src/agents/security.md +1 -0
  71. package/src/agents/skeptic.md +1 -0
  72. package/src/agents/spring-reviewer.md +1 -0
  73. package/src/agents/spring-security.md +1 -0
  74. package/src/agents/steward.md +1 -0
  75. package/src/agents/user-advocate.md +1 -0
  76. package/src/agents/visionary.md +1 -0
  77. package/src/bricks/ci-fix.md +12 -2
  78. package/src/kits/superpowers/missing.md +3 -0
  79. package/src/kits/superpowers/provider.md +7 -0
  80. package/src/kits/superpowers/rules.md +6 -0
  81. package/src/kits/superpowers.json +45 -0
  82. package/src/modes/foreman.md +18 -12
  83. package/src/skills/commit.md +4 -0
  84. package/src/skills/parallel-agents.md +2 -2
  85. package/src/skills/spring-boot/cloud.md +8 -2
  86. package/src/skills/spring-boot.md +3 -1
  87. package/src/skills/spring-testing.md +3 -3
  88. package/src/upstreams.json +1 -1
@@ -23,8 +23,9 @@ task that would block it to one background task worker, and keeps answering the
23
23
  proof commands, and the rule that it commits only in its own worktree, never pushes, and
24
24
  never merges outside its own branch; without a worktree, it never commits. It also covers
25
25
  the **Brief checklist** in the parallel-agents {{SKILL}}: read it before you write the brief. The worker explores, implements and tests in its own context, loading {{SKILLS}}
26
- as it needs them; after two repair rounds on its own red tests it returns red with its
27
- findings ({{DOCTRINE:bound-the-loop}}). Check its write set against every task in flight
26
+ as it needs them. It returns `STATUS: ERROR` when stuck and `STATUS: RELAY` when long.
27
+ The task-worker section of the developer spec sets both bounds
28
+ ({{DOCTRINE:bound-the-loop}}). Check its write set against every task in flight
28
29
  ({{DOCTRINE:one-writer-per-file}}).
29
30
  - **A worker may lead, inside its worktree.** It may dispatch read-only helpers (explorer,
30
31
  researcher) at any time, and writing sub-workers, each in its own worktree branched from
@@ -45,22 +46,27 @@ task that would block it to one background task worker, and keeps answering the
45
46
  goes back to the same worker.
46
47
  - **Keep a ledger; route follow-ups to the same worker.** Keep
47
48
  `{{DIR_NOTEBOOK}}/foreman-ledger.md`, one line per live task:
48
- `T3 · <task> · agent <id> · worktree <path> · <state>` (running, green, red, verifying,
49
- merged, dropped). Update it on every dispatch and return; re-read it after a compaction.
49
+ `T3 · <task> · agent <id> · worktree <path> · <state>` (running, green, red, relay,
50
+ verifying, merged, dropped). Update it on every dispatch and return; re-read it after a compaction.
50
51
  A correction, follow-up or verifier finding on a task goes to that task's worker, which
51
52
  resumes with its full context: by `SendMessage` on Claude Code (on OpenClaude only in its
52
53
  agent-teams mode); on OpenCode by the `subagent` tool with that worker's `sessionID` (V1:
53
54
  `task` with `task_id`); the same call without `sessionID` (V1: `task_id`) starts a fresh
54
55
  worker. Bob has no resume. Start a new worker only when the old one cannot be resumed —
55
56
  the host has no resume, or its context is spent — and hand it a note: the ask, the diff
56
- so far, the findings, what was tried. A new task gets a new worker by default. It may go
57
- to an idle worker whose own task is settled (merged or dropped) when it works in the
58
- same files or context and that worker's context is still light: the worker starts it on
59
- a new branch from the current BASE, with the full brief of *One worker per task* (BASE,
60
- write set, proof commands, the Brief checklist) and its write set checked against every
61
- task in flight. Light or heavy is a judgement call, not a number: write the reason (same
62
- files, context used) in the task's new ledger line. A worker never holds two live tasks;
63
- one whose context is heavy, or whose last task touched other code, retires.
57
+ so far, the findings, what was tried. On `STATUS: RELAY` (ledger `relay`), hand the task
58
+ off to a fresh worker on the same worktree and branch, with the handoff note. A hand-off
59
+ is no failure and no retry of yours. On `STATUS: ERROR` (ledger `red`) from the first
60
+ worker, hand off once with the findings. A task gets at most 2 hand-offs, counted in the
61
+ ledger line (`relay 2 of <id>`). When it needs a 3rd, or a worker that took a hand-off
62
+ returns `STATUS: ERROR`, take the task to the user. A new task gets a new
63
+ worker by default. It may go to an idle worker whose own task is settled (merged or
64
+ dropped) when it works in the same files and that worker's context is still under about
65
+ 100k tokens: the worker starts it on a new branch from the current BASE, with the full
66
+ brief of *One worker per task* (BASE, write set, proof commands, the Brief checklist) and
67
+ its write set checked against every task in flight. Write the reason (same files, context size) in the task's new ledger
68
+ line. Otherwise start a new worker. A worker never holds two live tasks; one past that
69
+ size, or whose last task touched other code, retires.
64
70
  - **Merge on proof you re-ran.** A worker hands back its branch, its local commits and any
65
71
  uncommitted work, with raw test + lint output. The parent re-runs those commands itself —
66
72
  a worker's log is the worker's account, not the state — merges the branch into its own,
@@ -19,6 +19,10 @@
19
19
  repo has no discernible convention, default to **Conventional Commits**:
20
20
  `type(scope): subject` (`feat`, `fix`, `docs`, `refactor`, `test`, `chore`…),
21
21
  with a `!` before the colon or a `BREAKING CHANGE:` footer for a breaking change.
22
+ **Pass the message as data the commit gate can read:** one quoted `-m '…'` for a
23
+ subject alone; for a body, under bash, `git commit -F - <<'EOF'` … `EOF` with the
24
+ delimiter quoted. Never `-m "$(cat <<'EOF' … EOF)"`: the gate cannot see inside a `$(`,
25
+ so it asks even on your own branch.
22
26
  5. **Commit on your own branch; never on the main branch ({{DOCTRINE:consent-before-push}}).**
23
27
  On your own branch, commit without asking once steps 1–4 hold. The main branch
24
28
  (`main`, `master`, the remote's default, and from a linked worktree the main checkout's branch) takes no direct
@@ -17,8 +17,8 @@
17
17
  - **Git state:** the ref to measure (`HEAD` or `origin/main`, after a fetch) and whether the diff is staged, on disk or committed; after a restart, `git log` before redoing "uncommitted" work — a stale ref redoes every finding.
18
18
  - **The whole write set:** every other call or literal site of a touched symbol, and its twin or generated copies, found by search before dispatch — a new field breaks the sites nobody listed.
19
19
  - **Scratch and load:** one scratch path per unit, at most cores/2 full suites at once, and default test concurrency where spawning is CPU-bound (Windows) — shared scratch clobbers, an oversubscribed suite fails as time-outs.
20
- - **Model tier:** you pick each unit's tier per task and name it in the brief. The strongest tier is your own current model (no override = inherit); step down, never up — intermediate for well-scoped writing, light for read-only search, counting and digests, your own for judgement (review, security, design) and for a retry after a failed attempt. Host with no lower tier, or model names unknown (OpenCode: set per machine): inherit. (Claude: opus > sonnet > haiku.)
21
- - **Long runs and the done gate:** a suite over about two minutes runs through the host's background mechanism and is waited on to completion before the unit ends — a foreground run dies at the tool's time-out. Where the project has a done gate (`.geneseed/done-checks.txt`), its checks run unpiped; output piped through `tail` or `grep` does not count.
20
+ - **Model tier:** you pick each unit's tier per task and name it in the brief. The strongest tier is your own current model (no override = inherit); step down, never up — intermediate for well-scoped writing, light for read-only search, counting and digests, your own for judgement (review, security, design) and for a retry after a failed attempt. The one exception: a model the user asked for (in chat, or as a pin in `agent-overrides.json`) may step up. On Claude Code, an install whose carrier says `Model tiers: auto` already runs each named seat at its tier (the judgement seats — reviewer, security, architect, performance and the kit reviewers — on your model; explorer, researcher and the council on a light model; the rest on an intermediate one), so pass no model for a named seat. Host with no lower tier, or model names unknown (OpenCode: set per machine): inherit. (Claude: opus > sonnet > haiku.)
21
+ - **Long runs and the done gate:** a suite over about two minutes runs through the host's background mechanism and is waited on to completion before the unit ends — a foreground run dies at the tool's time-out. Wait for the completion notice; never poll with a monitor or a sleep loop. While a task worker iterates, it runs the targeted test file with output to a file, and reads only the count line and the failing test names; the full suite runs once, at the end. A crew tester's logs stay raw. Where the project has a done gate (`.geneseed/done-checks.txt`), its checks run unpiped: a redirect to a file keeps the exit code, output piped through `tail` or `grep` does not count.
22
22
  - **Dependency installs:** whether a dependency install (`npm ci` or the stack's own) is allowed in a fresh worktree — a new worktree has none.
23
23
  - **Host shell limits:** one plain command per call, from the worktree as working directory or with absolute paths (`git -C`, `npm --prefix`), never opening with `cd <worktree> &&`; the isolated host may refuse or ask approval for heredocs, shell variables and compound commands, so prefer plain single commands unless the brief's proof commands need one; a multi-line commit message goes through `git commit -F <file>`, the file written with the edit tools. Paths outside the worktree (a scratch path) may be refused too: scratch files go through the file tools.
24
24
  - **One judge per copy:** when a global and a project install both run, a stand-down that lets one copy record for several is safe only for records every copy sees the same way (counts, child runs) — never for a verdict each copy reaches from its own config.
@@ -56,7 +56,13 @@ public Quote fetch(String id) { ... }
56
56
  public void sendNotification() { ... }
57
57
  ```
58
58
 
59
- Retry only idempotent calls. Circuit breaking: at the edge, use the Gateway `CircuitBreaker` filter; in-service, Resilience4j is still the common choice but its annotation and property names were not re-verified for Boot 4 here, so check its current docs before copying old config.
59
+ Retry only idempotent calls.
60
+ - `maxRetries` counts the retries after the first call: total attempts are 1 + `maxRetries`. The default is 3 retries, 1 second apart.
61
+ - Every exception is retried unless `includes` or `excludes` narrow the set; both also match nested causes.
62
+ - `@ConcurrencyLimit` caps simultaneous calls, like a pool size. It is not a rate limiter over time.
63
+ - Boot 4 no longer manages a Spring Retry version (`org.springframework.retry`): use the Framework annotations above, or declare the version yourself.
64
+
65
+ Framework 7 has no circuit breaker. At the edge, use the Gateway `CircuitBreaker` filter. In-service, Resilience4j publishes a Boot 4 module (`resilience4j-spring-boot4`), but its annotation and property names were not re-verified for Boot 4 here, so check its current docs before copying old config.
60
66
 
61
67
  ## Gateway
62
68
  Spring Cloud Gateway's reactive server uses the `spring.cloud.gateway.server.webflux.*` prefix in current releases (older guides show `spring.cloud.gateway.*`). Prefer YAML routes over the Java DSL for simple cases.
@@ -75,7 +81,7 @@ spring:
75
81
  - CircuitBreaker=usersBreaker
76
82
  ```
77
83
 
78
- The `CircuitBreaker` filter needs a Spring Cloud CircuitBreaker starter (for example the reactor-Resilience4j one) on the classpath. Gateway is WebFlux; do not put it in the same module as a Spring MVC service.
84
+ The `CircuitBreaker` filter needs a Spring Cloud CircuitBreaker starter (for example the reactor-Resilience4j one) on the classpath. Pick the starter for the server stack: `spring-cloud-starter-gateway-server-webflux` (reactive, the YAML above) or `spring-cloud-starter-gateway-server-webmvc` (Servlet, its properties under `spring.cloud.gateway.server.webmvc.*`). The WebFlux server cannot share a module with a Spring MVC service.
79
85
 
80
86
  ## Observability
81
87
  - Health: `/actuator/health/liveness` and `/actuator/health/readiness` are on by default (`management.endpoint.health.probes.enabled=false` disables them); point the Kubernetes probes at those, and permit `/actuator/health/**` in the security chain. Expose only what you need (`management.endpoints.web.exposure.include`), and protect the rest. `management.endpoint.health.show-components: always` with `show-details: when-authorized` tells an anonymous caller which component is down without leaking hosts or versions.
@@ -40,11 +40,13 @@ Removed or renamed in Boot 4 (do not write them): `@MockBean` and `@SpyBean`, `C
40
40
  | Security configuration | Spring Security 7: lambda DSL, `authorizeHttpRequests` |
41
41
  | `org.springframework.boot.test.autoconfigure.web.servlet.WebMvcTest` | `org.springframework.boot.webmvc.test.autoconfigure.WebMvcTest`: slices moved to per-technology modules; import from the new packages |
42
42
  | Jackson 2 (`com.fasterxml.jackson`) | Jackson 3 is the default (`tools.jackson`); Jackson 2 support is deprecated, `spring.jackson.use-jackson2-defaults` eases migration |
43
+ | Inject `ObjectMapper`, `@JsonComponent`, `@JsonMixin` | Inject the auto-configured `JsonMapper`; use `@JacksonComponent`, `@JacksonMixin`. Only the annotations keep their package (`com.fasterxml.jackson.annotation`) |
43
44
  | `@DynamicPropertySource` for container URLs | `@ServiceConnection` on the container |
44
45
  | MockMvc status assertions via Hamcrest | `MockMvcTester` (AssertJ) or `RestTestClient` with `@AutoConfigureRestTestClient` |
45
46
  | OpenFeign for declarative clients | `@HttpExchange` interfaces, registered with `@ImportHttpServices` |
46
47
  | Zipkin-only tracing setup | `spring-boot-starter-opentelemetry`, `management.opentelemetry.tracing.export.otlp.*` |
47
- | Starters assumed to pull everything | Boot 4 starters are modular; add the matching starter or test starter for each technology, and read the migration guide when a class vanishes |
48
+ | Starters assumed to pull everything | Boot 4 starters are modular; add the matching starter or test starter (`spring-boot-starter-<technology>-test`) for each technology, and read the migration guide when a class vanishes |
49
+ | `spring-boot-starter-web`, `-aop`, `-oauth2-*`; `flyway-core` or `liquibase-core` alone | `spring-boot-starter-webmvc`, `-aspectj`, `-security-oauth2-*`; `spring-boot-starter-flyway`, `-liquibase`. The old web and OAuth2 names are deprecated |
48
50
 
49
51
  ## Side files
50
52
  - **Web/REST** -> read [spring-boot/web.md](spring-boot/web.md) and follow it.
@@ -18,9 +18,9 @@ Target stack: Spring Boot 4, JUnit 5, Mockito, AssertJ.
18
18
  | Controller mapping, validation, serialization | `@WebMvcTest(X.class)` + `MockMvcTester` + `@MockitoBean` |
19
19
  | Repository queries, mappings | `@DataJpaTest` against the real database engine |
20
20
  | Whole app over HTTP | `@SpringBootTest(webEnvironment = RANDOM_PORT)` + `@AutoConfigureRestTestClient` |
21
- | Whole app, no server | `@SpringBootTest` + `@AutoConfigureMockMvc` |
21
+ | Whole app, no server | `@SpringBootTest` + `@AutoConfigureMockMvc` (`@SpringBootTest` alone no longer gives MockMvc) |
22
22
 
23
- Boot 4 split the test support into modules: slice annotations live in `org.springframework.boot.<module>.test.autoconfigure` (e.g. `org.springframework.boot.data.jpa.test.autoconfigure.DataJpaTest`); `@AutoConfigureTestDatabase` is in `org.springframework.boot.jdbc.test.autoconfigure`. Let the IDE import them. Use the matching test starter (`spring-boot-starter-data-jpa-test`, and so on) so each slice is on the classpath.
23
+ Boot 4 split the test support into modules: slice annotations live in `org.springframework.boot.<module>.test.autoconfigure` (e.g. `org.springframework.boot.data.jpa.test.autoconfigure.DataJpaTest`); `@AutoConfigureTestDatabase` is in `org.springframework.boot.jdbc.test.autoconfigure`. Let the IDE import them. Use the matching test starter (`spring-boot-starter-data-jpa-test`, and so on) so each slice is on the classpath. `@AutoConfigureRestTestClient` is in the `spring-boot-resttestclient` module, which `spring-boot-starter-webmvc-test` brings in; without that starter, add the module in test scope.
24
24
 
25
25
  ## Mocking beans
26
26
 
@@ -80,7 +80,7 @@ class ContainersConfig {
80
80
  class AuthorRepositoryTests { /* ... */ }
81
81
  ```
82
82
 
83
- - Testcontainers 2 containers are non-generic and live in per-module packages (`org.testcontainers.postgresql.PostgreSQLContainer`, artifact `testcontainers-postgresql`).
83
+ - Testcontainers 2 containers are non-generic and live in per-module packages (`org.testcontainers.postgresql.PostgreSQLContainer`, artifact `testcontainers-postgresql`). With them, add `spring-boot-testcontainers` and, for `@Testcontainers`/`@Container`, `org.testcontainers:testcontainers-junit-jupiter`, all in test scope.
84
84
  - The default `replace = NON_TEST` already keeps a `@ServiceConnection` data source; `replace = NONE` is only needed for a non-test `DataSource` you defined yourself.
85
85
  - Where `@ServiceConnection` has no support for the container, use `@Testcontainers` + `@Container static` + `@DynamicPropertySource` to register the properties.
86
86
  - Share one container configuration across tests so the context cache is reused.
@@ -58,7 +58,7 @@
58
58
  },
59
59
  "piomin-spring-boot": {
60
60
  "repo": "piomin/claude-ai-spring-boot",
61
- "commit": "647d768a5d333e15b6a50dc0e8fc41e41a330e92",
61
+ "commit": "cbd289d4adadef06bd2ac118294319a2f49f79ee",
62
62
  "author": "Piotr Mińkowski",
63
63
  "license": "Apache-2.0",
64
64
  "licenseFile": "licenses/piomin-spring-boot.txt"