ultra-builder-pro-cli 3.0.2 → 5.0.1

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 (42) hide show
  1. package/.ultra-template/contexts/TEMPLATE.md +6 -0
  2. package/CHANGELOG.md +83 -1
  3. package/README.md +126 -36
  4. package/README.zh-CN.md +440 -0
  5. package/adapters/_shared/host-profile.cjs +3 -151
  6. package/adapters/_shared/plugin-core.cjs +57 -0
  7. package/adapters/_shared/runtime-assets.cjs +64 -12
  8. package/adapters/codex.js +3 -6
  9. package/bin/delegate-worker.cjs +2 -345
  10. package/bin/delegate.cjs +3 -530
  11. package/docs/ARCHITECTURE.md +59 -18
  12. package/docs/ARTIFACT-AUTHORITY.md +21 -5
  13. package/docs/DECISIONS.md +31 -10
  14. package/docs/PLUGIN-ISOLATION-CONTRACT.md +28 -12
  15. package/docs/RUNTIME-COMPAT-MATRIX.md +106 -26
  16. package/docs/SKILL-AUTHORING.md +15 -3
  17. package/docs/WORKFLOW-LIFECYCLE.md +83 -29
  18. package/docs/evals/ultra-5-0-journey-results.md +62 -0
  19. package/docs/evals/ultra-5-0-live-acp-conformance.md +93 -0
  20. package/docs/evals/ultra-5-0-route-journeys.md +145 -0
  21. package/package.json +2 -1
  22. package/skills/ultra-change/SKILL.md +30 -4
  23. package/skills/ultra-delegate/SKILL.md +80 -7
  24. package/skills/ultra-delegate/patches/zcode-acp-0.1.0-zcode-0.16.5.patch +184 -0
  25. package/skills/ultra-delegate/references/delegation-contract.md +60 -1
  26. package/skills/ultra-delegate/scripts/acp-conformance.cjs +287 -0
  27. package/skills/ultra-delegate/scripts/acp-driver.cjs +762 -0
  28. package/skills/ultra-delegate/scripts/agent-execution-driver.cjs +320 -0
  29. package/skills/ultra-delegate/scripts/cli-driver.cjs +61 -0
  30. package/skills/ultra-delegate/scripts/delegate-worker.cjs +328 -0
  31. package/skills/ultra-delegate/scripts/delegate.cjs +690 -0
  32. package/skills/ultra-delegate/scripts/host-profile.cjs +267 -0
  33. package/skills/ultra-deliver/SKILL.md +13 -0
  34. package/skills/ultra-dev/SKILL.md +18 -2
  35. package/skills/ultra-init/SKILL.md +12 -0
  36. package/skills/ultra-plan/SKILL.md +21 -3
  37. package/skills/ultra-research/SKILL.md +33 -5
  38. package/skills/ultra-status/SKILL.md +37 -7
  39. package/skills/ultra-status/references/route-contract.md +100 -0
  40. package/skills/ultra-test/SKILL.md +16 -1
  41. /package/{adapters/_shared → skills/ultra-delegate/scripts}/delegate-profiles/kimi-read-only.md +0 -0
  42. /package/{adapters/_shared → skills/ultra-delegate/scripts}/delegate-profiles/kimi-write.md +0 -0
@@ -0,0 +1,100 @@
1
+ # Route Contract
2
+
3
+ This reference defines the owner-visible navigation surface shared by every public
4
+ Ultra Skill. A Route Card is a derived observation from canonical project files, Git,
5
+ and the current run. It is not workflow authority, a stage marker, or permission to
6
+ continue.
7
+
8
+ ## Route Card
9
+
10
+ Show a compact `UBP Route Card` at entry, at a material checkpoint, when pausing or
11
+ blocked, at the observable exit, and before moving to another public Skill. Keep an
12
+ unchanged field short, but never omit an unresolved decision or unfinished obligation.
13
+
14
+ | Field | Meaning |
15
+ |---|---|
16
+ | `Route` | The current public Skill, or `idle` when no workflow is active |
17
+ | `Mode` | One mode declared by that Skill |
18
+ | `Why` | The current problem or evidence gap |
19
+ | `Outcome` | The owner-visible result this route is trying to produce |
20
+ | `Accepted boundary` | Accepted scope, exclusions, and authority that cover this run |
21
+ | `Position` | Artifact-derived navigation such as selected reports or remaining tasks |
22
+ | `Exit condition` | Observable evidence or owner decision that ends this mode |
23
+ | `Delta` | Material facts changed since the prior card |
24
+ | `Reality` | Live path, verification state, fakes, limitations, and unknowns |
25
+ | `Decision needed` | One material owner decision, or `none` |
26
+ | `Next` | The smallest covered continuation or one recommended public route |
27
+ | `Not done` | Obligations left at a pause or exit |
28
+
29
+ `Position` may report exact mechanical observations such as selected lenses with
30
+ reports, dependency-ready tasks, or an ACP session waiting for permission. Never turn a
31
+ count, validator result, finding total, or transport event into semantic progress or a
32
+ numeric completion score.
33
+
34
+ ## Material checkpoints
35
+
36
+ Stop and obtain the Owner's current decision before any of these changes or effects:
37
+
38
+ 1. a new or reduced North Star, specification, or Change commitment;
39
+ 2. a material delta to an active Change outcome, acceptance, non-goal, or public seam;
40
+ 3. evidence that materially changes the feasible direction or product boundary;
41
+ 4. wider Execution Grant scope, effects, budget, expiry, or Agent/provider topology;
42
+ 5. transfer of the canonical writer;
43
+ 6. unresolved P0/P1, risk acceptance, or the same root cause surviving three failed fixes;
44
+ 7. finalization, archive, commit, push, tag, publication, deployment, real installation,
45
+ credentials, provider spend, or another external or irreversible effect.
46
+
47
+ Naming, internal design, ordinary technical choices, task decomposition, covered local
48
+ edits and tests, and non-material evidence reconciliation remain model-owned inside the
49
+ accepted boundary. Do not ask for confirmation when `Decision needed` is `none`.
50
+
51
+ ## Continuation
52
+
53
+ Public modes belong to one of three invocation classes; model-only disciplines form a
54
+ fourth class:
55
+
56
+ | Invocation class | Owner entry | Model continuation |
57
+ |---|---:|---:|
58
+ | `explicit-entry` | yes | no |
59
+ | `grant-continuable` | yes | only under a valid Execution Grant covering that mode |
60
+ | `router` | yes | no; the Owner invokes read-only diagnosis explicitly |
61
+ | `model-discipline` | no | only as an internal method of a covered workflow |
62
+
63
+ Host metadata is only a coarse discovery bit. Before continuing, the Skill itself must
64
+ verify its selected mode, the same active `change_id`, the valid Execution Grant,
65
+ allowed local effects, stop conditions, and any material checkpoint. A coarse Host bit
66
+ never permits an explicit-entry mode. Without a covering grant, stop with a Route Card
67
+ and recommend the exact explicit route; never invoke another public Skill merely because
68
+ its files appear ready.
69
+
70
+ Within an accepted Change and valid grant, show the Route Card before each public Skill
71
+ transition and continue without a confirmation question only when the next mode and all
72
+ effects are covered. A material checkpoint still pauses. `Next` is either `continue:
73
+ <route>/<mode>` when covered or `recommend: <route>/<mode>` when owner invocation is
74
+ required.
75
+
76
+ Execution events such as tool calls, plans, usage, permission requests, worker output,
77
+ ACP updates, and process termination are transport observations. They may populate
78
+ `Position`, `Delta`, or `Reality`; they never decide acceptance, task completion, or the
79
+ next semantic route.
80
+
81
+ ## Recovery and degraded diagnosis
82
+
83
+ On a fresh session or Host, reconstruct the card from the accepted North Star, Project
84
+ Brief, active Change authority, task ledger and matching contexts, decisions, evidence,
85
+ Git, and current installation/runtime observations. Read the closing `## Resume Note`
86
+ for only the current work package. Do not rerun completed Research or create a route
87
+ ledger to remember where the conversation stopped.
88
+
89
+ If authority is absent, ambiguous, stale, malformed, or temporarily unreadable, stop
90
+ semantic writes but keep read-only diagnosis reachable. Report the exact fact that
91
+ prevents a normal route and the cheapest safe repair, retry, cancel, abandon, or fresh
92
+ session path. Never describe a blocked route as an endless in-progress workflow.
93
+
94
+ Owner revisions remain recoverable: edit a draft in place, supersede an accepted
95
+ decision with a new owner record, and preserve Git/history. Status may summarize that
96
+ history on request, but no event log becomes a second source of truth.
97
+
98
+ When context pressure threatens reliable continuation, finish the nearest safe write,
99
+ refresh the canonical context's Resume Note, show a final card, and recommend a fresh
100
+ session. Context size is an observation, not a hard semantic gate.
@@ -9,6 +9,18 @@ This is the one end-of-change sensor allowed by PHILOSOPHY C4: an export cannot
9
9
  called orphaned while a later task may still wire it. It is a terminal sensor, never a gate;
10
10
  the owner decides what to fix or accept.
11
11
 
12
+ ## Route modes
13
+
14
+ | Mode | Entry | Observable exit | Next |
15
+ |---|---|---|---|
16
+ | `audit` | Explicit invocation or a valid Execution Grant after every matching task has current completion evidence | One current whole-Change report and aggregate review preserve all six audit areas, findings, omissions, identities, and dispositions | Recommend `ultra-deliver/reconcile` when current, or the owner-selected responsible route |
17
+ | `reaudit` | Explicit invocation or a valid Execution Grant after an owner-dispositioned repair invalidates the prior report | A fresh report binds the repaired Change snapshot without hiding earlier findings | Recommend `ultra-deliver/reconcile` when current, or another bounded owner disposition |
18
+
19
+ Follow `../ultra-status/references/route-contract.md`. Show a Route Card at audit entry,
20
+ before asking for finding disposition, and at the report exit. Audit counts and review
21
+ events populate `Position` or `Reality`; they never choose repair, convergence, or risk
22
+ acceptance. Test does not call its own repair route.
23
+
12
24
  ## Before you start
13
25
 
14
26
  1. If model-selected, verify the live execution grant in `../ultra-change/references/execution-grant.md` — a current session-local activation or a stably verified durable work-package grant; without either, stop.
@@ -209,7 +221,9 @@ disposition as separate fields. No validator result is itself semantic acceptanc
209
221
 
210
222
  Present the findings by consequence and recommend the highest-leverage response. The
211
223
  owner may fix all, fix selected findings, accept recorded risk, or reduce scope through
212
- the normal REDUCTION decision. Recommend the next capability; do not invoke it.
224
+ the normal REDUCTION decision. Recommend the next capability. Continue only when the
225
+ owner disposition or existing grant covers its exact mode and effects, and show the
226
+ transition Route Card first; Test itself never manufactures that coverage.
213
227
 
214
228
  ## When the owner decides
215
229
 
@@ -225,5 +239,6 @@ an observation, not permission to auto-fix, weaken a test or silently change sco
225
239
  - `../ultra-review/SKILL.md` — read for aggregate Change review after the six-area audit.
226
240
  - `../ultra-plan/references/task-evidence-v2.md` — canonical current task evidence and task-review contract.
227
241
  - `../ultra-change/references/execution-grant.md` — read only for grant-activated continuation.
242
+ - `../ultra-status/references/route-contract.md` — owner-visible route, checkpoint, and recovery output.
228
243
  - `../ultra-think/references/autonomy-boundary.md` — read when a proposed response
229
244
  would make an existing specification commitment stop holding.