القائمة
Configure SDLC model tiers. Interactive TUI via /sdlc-choose, quick switch via /sdlc-tier.
التثبيت باستخدام Codex أو Claude انسخ هذا Prompt والصقه في Codex أو Claude أو مساعد آخر ليراجع صفحة Skill ويثبّتها لك.
استنادا إلى تصنيف SOC المهني
Execute tasks with pre/post drift checks and verification gates. Use after sdlc-plan to implement with accuracy guarantees.
Final verification against spec with full evidence chain. Use after sdlc-execute to confirm feature is complete.
Break specs into ordered tasks with acceptance criteria. Use after sdlc-spec to create executable task breakdown.
Create feature specs through collaborative brainstorming. Use when starting new features, writing requirements, or defining acceptance criteria.
| name | sdlc-choose |
| description | Configure SDLC model tiers. Interactive TUI via /sdlc-choose, quick switch via /sdlc-tier. |
This skill documents extension commands. The TUI is in
extensions/sdlc-choose.ts.
Interactive TUI for model configuration:
/sdlc-choose # Open menu
/sdlc-choose show # Show current
/sdlc-choose high # Quick set tier
/sdlc-choose list # List available models
Quick tier switch:
/sdlc-tier # Show current
/sdlc-tier high # Set tier
/sdlc-tier budget
Direct model switch:
/sdlc-model local-llm/alibaba/qwen-max
/sdlc-model anthropic/claude-sonnet-4-6
Extension auto-switches model when SDLC skills invoked:
| Skill | Config Key | Purpose |
|---|---|---|
/skill:sdlc-spec | models.{tier}.spec | Reasoning, planning |
/skill:sdlc-plan | models.{tier}.plan | Task breakdown |
/skill:sdlc-execute | models.{tier}.execute | Coding |
/skill:sdlc-verify | models.{tier}.verify | Review, testing |
.sdlc/config.json in project root:
{
"tier": "medium",
"models": {
"high": {
"spec": "anthropic/claude-opus-4-7",
"plan": "anthropic/claude-opus-4-6",
"execute": "anthropic/claude-sonnet-4-6",
"verify": "anthropic/claude-sonnet-4-6"
},
"medium": {
"spec": "google/gemini-2.5-pro",
"plan": "google/gemini-2.5-pro",
"execute": "openai/gpt-4o",
"verify": "openai/gpt-4o-mini"
},
"budget": {
"spec": "deepseek/deepseek-r1",
"plan": "deepseek/deepseek-r1",
"execute": "deepseek/deepseek-coder-v3",
"verify": "google/gemini-flash"
}
}
}
Use full provider/model ID to avoid conflicts with proxies:
# Format: provider/model-id
local-llm/alibaba/qwen-max
anthropic/claude-sonnet-4-6
google/gemini-2.5-pro
openai/gpt-4o
Run /sdlc-choose list or pi --list-models to see available IDs.
export SDLC_TIER=budget
SDLC_TIER=high pi "/skill:sdlc-spec"
Shows SDLC: {tier} when .sdlc/config.json exists.