umadev 1.0.14 → 1.0.16

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 (2) hide show
  1. package/README.md +16 -8
  2. package/package.json +8 -8
package/README.md CHANGED
@@ -1,9 +1,12 @@
1
1
  # UmaDev
2
2
 
3
- > **A governance rail around your AI coding base** — drives the Claude Code /
4
- > Codex / OpenCode you already logged into through a 9-phase commercial
5
- > delivery pipeline. The base writes the code; UmaDev runs the process.
6
- > **No API key needed.** Early-stage and governance-first.
3
+ > **UmaDev: A coding agent that works like a real dev team** — product manager,
4
+ > architect, UI/UX designer, frontend, backend, QA, security, and DevOps, each doing
5
+ > its own specialty on a shared blackboard, borrowing the Claude Code / Codex /
6
+ > OpenCode you already logged into as the brain. A **coordinator** seat schedules the
7
+ > team and enforces the gates (it routes, plans, and gates — it is not the headline);
8
+ > the base writes the code. You don't hire a director — you hire the whole team.
9
+ > **No API key needed.**
7
10
 
8
11
  ## Install
9
12
 
@@ -33,10 +36,15 @@ umadev report # emit UD-EVID-004 compliance map
33
36
  ## Why this exists
34
37
 
35
38
  UmaDev is **not** an LLM client. It does not call any AI API.
36
- Instead it **drives** one of the three first-class base CLIs you already
37
- use (`claude`, `codex`, `opencode`) through a deterministic 9-phase
38
- pipeline — the brain stays in the base. Wider model coverage is the base's
39
- job (route it to a third-party / local model), not a new UmaDev driver:
39
+ Instead it convenes a development team — eight role specialists that plan,
40
+ build, review, and sign off like a real team — over one of the three
41
+ first-class base CLIs you already use (`claude`, `codex`, `opencode`); the
42
+ brain stays in the base. Wider model coverage is the base's job (route it to
43
+ a third-party / local model), not a new UmaDev driver.
44
+
45
+ The coordinator routes each request: a chat stays chat, a one-line edit takes
46
+ the fast path, and only a full product requirement expands into the team's
47
+ deepest play — the deterministic commercial delivery chain:
40
48
 
41
49
  ```
42
50
  research → docs → ⏸ docs_confirm → spec → frontend → ⏸ preview_confirm → backend → quality → delivery
package/package.json CHANGED
@@ -1,7 +1,7 @@
1
1
  {
2
2
  "name": "umadev",
3
- "version": "1.0.14",
4
- "description": "A project-director Agent for AI coding hosts — drives your logged-in Claude Code / Codex through a 9-phase commercial delivery pipeline with governance. No API key needed.",
3
+ "version": "1.0.16",
4
+ "description": "UmaDev: A coding agent that works like a real dev team — PM, architect, designer, frontend, backend, QA, security, DevOps — borrowing your logged-in Claude Code / Codex / OpenCode brain to turn one requirement into a shippable, commercial-grade app, coordinated by a scheduling seat. No API key needed.",
5
5
  "keywords": [
6
6
  "ai",
7
7
  "claude-code",
@@ -37,11 +37,11 @@
37
37
  "node": ">=18"
38
38
  },
39
39
  "optionalDependencies": {
40
- "@umacloud/cli-darwin-arm64": "1.0.14",
41
- "@umacloud/cli-darwin-x64": "1.0.14",
42
- "@umacloud/cli-linux-x64": "1.0.14",
43
- "@umacloud/cli-linux-arm64": "1.0.14",
44
- "@umacloud/cli-win32-x64": "1.0.14",
45
- "@umacloud/knowledge": "1.0.14"
40
+ "@umacloud/cli-darwin-arm64": "1.0.16",
41
+ "@umacloud/cli-darwin-x64": "1.0.16",
42
+ "@umacloud/cli-linux-x64": "1.0.16",
43
+ "@umacloud/cli-linux-arm64": "1.0.16",
44
+ "@umacloud/cli-win32-x64": "1.0.16",
45
+ "@umacloud/knowledge": "1.0.16"
46
46
  }
47
47
  }