ultra-builder-pro-cli 0.5.3 → 0.6.0

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 (47) hide show
  1. package/CHANGELOG.md +31 -6
  2. package/README.md +14 -16
  3. package/adapters/_shared/codex-assets.cjs +32 -151
  4. package/adapters/_shared/frontmatter.cjs +2 -2
  5. package/adapters/_shared/handbook.cjs +3 -3
  6. package/adapters/_shared/path-rewrite.cjs +0 -3
  7. package/adapters/_shared/runtime-assets.cjs +11 -4
  8. package/adapters/_shared/tests/path-rewrite.test.cjs +1 -3
  9. package/adapters/_shared/tests/resolve-target.test.cjs +1 -11
  10. package/adapters/_shared/tests/runtime-assets.test.cjs +5 -6
  11. package/adapters/opencode.js +1 -6
  12. package/adapters/tests/claude.test.cjs +1 -4
  13. package/adapters/tests/codex.test.cjs +5 -6
  14. package/adapters/tests/opencode.test.cjs +1 -2
  15. package/bin/install.js +6 -9
  16. package/bin/orchestrator.js +11 -4
  17. package/docs/AGENT-CONTEXT.md +4 -5
  18. package/docs/ARCHITECTURE.md +3 -4
  19. package/docs/LEGACY-HERMES.md +5 -81
  20. package/docs/PLAN.zh-CN.md +35 -42
  21. package/docs/ROADMAP.md +3 -3
  22. package/docs/RUNTIME-COMPAT-MATRIX.md +61 -60
  23. package/mcp-server/lib/pricing.cjs +0 -6
  24. package/mcp-server/lib/state-ops.cjs +7 -0
  25. package/mcp-server/lib/telemetry.cjs +4 -0
  26. package/mcp-server/tests/server-contract.test.cjs +17 -0
  27. package/mcp-server/tests/state-ops.test.cjs +23 -0
  28. package/mcp-server/tests/telemetry.test.cjs +14 -0
  29. package/orchestrator/dispatch-rules.cjs +15 -4
  30. package/orchestrator/tests/daemon.test.cjs +13 -12
  31. package/orchestrator/tests/dispatch-rules.test.cjs +12 -10
  32. package/package.json +3 -5
  33. package/skills/ultra-init/SKILL.md +0 -1
  34. package/skills/ultra-verify/SKILL.md +22 -41
  35. package/skills/ultra-verify/evals/evals.json +22 -24
  36. package/skills/ultra-verify/references/confidence-system.md +15 -32
  37. package/skills/ultra-verify/references/cross-verify-modes.md +14 -59
  38. package/skills/ultra-verify/scripts/verify_wait.py +44 -81
  39. package/spec/mcp-tools.yaml +3 -3
  40. package/spec/schemas/skill-manifest.schema.json +2 -2
  41. package/spec/schemas/state-db.sql +2 -2
  42. package/adapters/_shared/md-to-toml.cjs +0 -54
  43. package/adapters/_shared/tests/md-to-toml.test.cjs +0 -48
  44. package/adapters/gemini.js +0 -204
  45. package/adapters/tests/gemini.test.cjs +0 -112
  46. package/skills/gemini-collab/SKILL.md +0 -51
  47. package/skills/gemini-collab/evals/evals.json +0 -42
@@ -1,51 +0,0 @@
1
- ---
2
- name: gemini-collab
3
- description: Ask Gemini CLI for an explicitly requested, read-only second opinion on a bounded review, diagnosis, architecture, or comparison question.
4
- ---
5
-
6
- # Gemini Collaboration
7
-
8
- Claude Code remains primary and owns scope, evidence, edits, verification, and the final answer.
9
- Gemini is an untrusted read-only advisor. Use this skill only when the user explicitly requests a
10
- Gemini perspective.
11
-
12
- ## Preconditions
13
-
14
- 1. Confirm `gemini --version` and authentication.
15
- 2. Fix the workspace, files or diff range, question, evidence standard, and expected response shape.
16
- 3. Write the primary analysis first when independence matters.
17
- 4. Never send secrets, unrelated files, or an unbounded home directory.
18
-
19
- ## Modes
20
-
21
- - `review`: findings against an exact diff or file set.
22
- - `understand`: a bounded architecture or flow question.
23
- - `opinion`: one decision with named constraints and alternatives.
24
- - `compare`: independent answers to the same question.
25
- - `free`: another tightly scoped read-only prompt.
26
-
27
- ## Safe invocation
28
-
29
- Create a session directory under `.ultra/collab/`, then use Gemini's plan approval mode:
30
-
31
- ```bash
32
- SESSION_PATH=".ultra/collab/$(date +%Y%m%d-%H%M%S)-gemini-<mode>"
33
- mkdir -p "${SESSION_PATH}"
34
- gemini --approval-mode plan \
35
- --output-format text \
36
- -p "<bounded prompt>" \
37
- > "${SESSION_PATH}/gemini-output.md" \
38
- 2> "${SESSION_PATH}/gemini-error.log"
39
- ```
40
-
41
- Do not use auto-edit, YOLO, or permission bypass.
42
-
43
- ## Synthesis
44
-
45
- 1. Read the output only after the process exits and the file is non-empty.
46
- 2. Verify consequential claims against the current checkout, tests, runtime, or primary docs.
47
- 3. Separate agreement, useful dissent, and unsupported assertions.
48
- 4. Return one Claude Code-owned conclusion; do not paste an unreviewed advisor transcript.
49
-
50
- If Gemini is missing, unauthenticated, times out, or returns empty output, report the degraded path
51
- and continue with primary evidence. Never block the user's task solely on advisor failure.
@@ -1,42 +0,0 @@
1
- {
2
- "skill_name": "gemini-collab",
3
- "evals": [
4
- {
5
- "id": 1,
6
- "prompt": "/gemini-collab review",
7
- "expected_output": "Claude creates a session directory, invokes gemini CLI to review current changes, reads the output file, and presents a synthesis report",
8
- "assertions": [
9
- "Creates a session directory under .ultra/collab/ with gemini-review in the name",
10
- "Uses 'gemini -p' with --yolo flag to invoke Gemini CLI",
11
- "Does NOT use '-o text' flag (Gemini outputs to stdout by default)",
12
- "Redirects output to a file (> output.md)",
13
- "Uses Read tool to read the output file",
14
- "Presents a synthesis with Consensus and Divergent Views sections"
15
- ]
16
- },
17
- {
18
- "id": 2,
19
- "prompt": "/gemini-collab compare 'Should we use Redis or Memcached for our caching layer?'",
20
- "expected_output": "Claude answers independently first, then sends the same question to Gemini, reads output, and presents a comparative synthesis",
21
- "assertions": [
22
- "Claude writes its own answer BEFORE invoking Gemini",
23
- "Sends the question to Gemini without revealing Claude's position",
24
- "Creates session directory under .ultra/collab/",
25
- "Reads Gemini output with Read tool",
26
- "Presents both analyses side by side with synthesis"
27
- ]
28
- },
29
- {
30
- "id": 3,
31
- "prompt": "/gemini-collab understand",
32
- "expected_output": "Claude invokes Gemini in sandbox mode to analyze the project, reads output, and synthesizes with its own understanding",
33
- "assertions": [
34
- "Uses --sandbox flag for read-only analysis",
35
- "Uses gemini -p with a project analysis prompt",
36
- "Installs from @google/gemini-cli (not @anthropic-ai/gemini-cli)",
37
- "Redirects to file, reads with Read tool",
38
- "Produces a unified project understanding report"
39
- ]
40
- }
41
- ]
42
- }