opencode/orchestrator.yaml at main · dotbrains/opencode · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
# Orchestrator Configuration for opencode

# ── Model Tiers ─────────────────────────────────────
#
# Cost optimization: route different complexity prompts to different models.
# Higher tiers = more capable but more expensive.

model-tiers:
  planning:
    model: "opencode-go/qwen3.6-plus"
    teams: "planning"
    description: "1M context - free, great for architecture"

  engineering:
    model: "opencode/minimax-m2.5-free"
    teams: "engineering"
    description: "200k context - free, strong coding"

  validation:
    model: "opencode/minimax-m2.5-free"
    teams: "validation"
    description: "200k context - free, good for reviews"

# ── Team Definitions ───────────────────────────────

teams:
  engineering:
    lead: engineering-lead
    workers:
      - frontend-dev
      - backend-dev
      - devops-specialist
    description: "Full-stack engineering — implements features"
    model-tier: engineering

  planning:
    lead: planning-lead
    workers:
      - planning-agent
      - plan-reviewer
    description: "Architecture and planning — plans and critiques"
    model-tier: planning

  validation:
    lead: validation-lead
    workers:
      - reviewer
      - security-reviewer
      - qa-engineer
    description: "Quality assurance — reviews, security, testing"
    model-tier: validation

# ── Routing Rules ───────────────────────────────────

routing:
  new-feature:
    - planning
    - engineering
    - validation
  bug-fix:
    - engineering
    - validation
  code-review:
    - validation
  architecture:
    - planning
  security-audit:
    - validation
  quick-lookup:
    - engineering

# ── Preprocessing ───────────────────────────────────

preprocess:
  # Requirements clarifier is always called first for these task types
  always_clarify:
    - new_feature
    - unclear_requirements
    - ambiguous_task

  # For these, go straight to relevant team
  skip_clarify:
    - bug_fix
    - code_review
    - quick_lookup

# ── Worktree Configuration ───────────────────────────────

worktrees:
  # Global location for worktrees (outside project directory)
  location: "~/.config/superpowers/worktrees"
  # Create worktrees for multi-team workflows
  enabled: true
  # Auto-cleanup after completion
  auto-cleanup: true

# ── Dynamic Skill Injection ───────────────────────────────

# Map task patterns to skills that should be injected
skill-injection:
  # Trigger phrases that indicate a skill should be used
  triggers:
    # Keywords → recommended skill
    "build me": brainstorming
    "create": brainstorming
    "design": brainstorming
    "implement": test-driven-development
    "add feature": test-driven-development
    "fix": systematic-debugging
    "debug": systematic-debugging
    "broken": systematic-debugging
    "security": security-and-hardening
    "auth": security-and-hardening
    "complex": planning-and-task-breakdown
    "refactor": code-simplification
    "review": code-review-and-quality
    "audit": mt-audit

  # Per-workflow default skills
  workflow-defaults:
    new-feature:
      - brainstorming  # Explore idea before planning
      - writing-plans   # Create implementation plan
      - test-driven-development  # Implement with tests
      - code-review-and-quality  # Review before merge
      - finishing-a-development-branch  # Merge/cleanup

    bug-fix:
      - systematic-debugging  # Find root cause
      - verification-before-completion  # Verify fix

    security-audit:
      - security-and-hardening  # Focused security review
      - mt-audit

    code-review:
      - code-review-and-quality