baro-ai 0.7.4 → 0.8.0

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 (2) hide show
  1. package/README.md +46 -10
  2. package/package.json +1 -1
package/README.md CHANGED
@@ -28,41 +28,77 @@ baro "Add authentication with JWT and role-based access control"
28
28
  # Use OpenAI for planning
29
29
  baro --planner openai "Add WebSocket support"
30
30
 
31
+ # Limit parallelism to 3 concurrent stories
32
+ baro --parallel 3 "Refactor database layer"
33
+
34
+ # Set story timeout to 5 minutes
35
+ baro --timeout 300 "Add unit tests"
36
+
37
+ # Force a specific model for all phases
38
+ baro --model opus "Complex architecture redesign"
39
+
40
+ # Disable model routing (use opus everywhere)
41
+ baro --no-model-routing "Build entire app"
42
+
31
43
  # Resume interrupted execution
32
44
  baro --resume
33
45
 
34
46
  # Specify working directory
35
- baro --cwd ~/projects/myapp "Add unit tests"
47
+ baro --cwd ~/projects/myapp "Add REST API"
36
48
  ```
37
49
 
38
50
  ## How it works
39
51
 
40
- 1. **Plan** — Claude explores your codebase and generates a dependency graph of user stories
41
- 2. **Review** — You review the plan, accept or quit
42
- 3. **Execute** — Stories run in parallel on a feature branch, each with its own Claude agent
43
- 4. **Review Agent** — After each level completes, a review agent checks the work against acceptance criteria and creates fix stories if needed
44
- 5. **Finalize** — Creates a GitHub PR with summary when all stories complete
52
+ 1. **Plan** — Claude (Opus) explores your codebase and generates a dependency graph of user stories
53
+ 2. **Review** — You review the plan, refine with feedback, accept or quit
54
+ 3. **Execute** — Stories run in parallel on a feature branch, each with its own Claude agent (Sonnet)
55
+ 4. **Review Agent** — After each level, a review agent (Haiku) checks work against acceptance criteria and creates fix stories if needed
56
+ 5. **Finalize** — Runs build verification and creates a GitHub PR with full summary
45
57
 
46
58
  ## Features
47
59
 
48
60
  - **Parallel execution** — independent stories run simultaneously, respecting dependency order
49
61
  - **DAG engine** — topological sort with level grouping, cycle detection
62
+ - **Model routing** — Opus for planning, Sonnet for execution, Haiku for review (configurable)
50
63
  - **Live TUI** — dashboard with story status, live agent logs, DAG view, stats
51
- - **Review agent** — automated code review between levels, auto-fix with retry
64
+ - **Review agent** — automated code review between levels with build detection and auto-fix
65
+ - **Plan refinement** — press `r` on review screen to give feedback and regenerate the plan
52
66
  - **Build detection** — auto-detects project type (Cargo, npm, Go, Python, Make) and runs builds during review
53
67
  - **Git coordination** — mutex-protected commits, auto-push with retry, pull --rebase, conflict detection
54
68
  - **Branch per run** — creates `baro/<name>` branch, keeps main clean
55
69
  - **Resume** — detects `prd.json` and resumes incomplete executions
56
- - **PR creation** — creates GitHub PR via `gh` CLI when done
70
+ - **PR creation** — creates GitHub PR with stories table, stats, time saved, and review summary
71
+ - **Configurable parallelism** — `--parallel N` to limit concurrent story execution
72
+ - **Story timeout** — `--timeout SECONDS` kills stuck agents (default: 10 minutes)
73
+ - **Time saved** — shows parallel speedup vs sequential execution
74
+ - **System notifications** — terminal bell + OS notification (macOS/Linux) when done
57
75
  - **Retry logic** — failed stories retry automatically (configurable per story)
58
- - **Claude + OpenAI** — Claude as default planner/executor, OpenAI as alternative planner
76
+
77
+ ## Options
78
+
79
+ ```
80
+ baro [goal] [options]
81
+
82
+ Arguments:
83
+ goal Project goal (opens welcome screen if omitted)
84
+
85
+ Options:
86
+ --planner <name> Planner: claude or openai (default: claude)
87
+ --model <name> Override model for all phases: opus, sonnet, haiku
88
+ --no-model-routing Use opus for everything (disables routing)
89
+ --parallel <N> Max concurrent stories, 0 = unlimited (default: 0)
90
+ --timeout <seconds> Story timeout in seconds (default: 600)
91
+ --resume Resume from existing prd.json
92
+ --cwd <path> Working directory (default: current)
93
+ -h, --help Print help
94
+ ```
59
95
 
60
96
  ## Requirements
61
97
 
62
98
  - [Claude CLI](https://docs.anthropic.com/en/docs/claude-cli) installed and authenticated
63
99
  - macOS (arm64/x64) or Linux (x64/arm64)
64
100
  - Node.js 18+ (only if using `--planner openai`)
65
- - `gh` CLI (optional, for PR creation)
101
+ - `gh` CLI (optional, for automatic PR creation)
66
102
 
67
103
  ## Architecture
68
104
 
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "baro-ai",
3
- "version": "0.7.4",
3
+ "version": "0.8.0",
4
4
  "description": "Autonomous parallel coding - plan and execute with AI",
5
5
  "type": "module",
6
6
  "bin": {