feat(profiles): add model profiles for GSD agents
Adds switchable model profiles (quality/balanced/budget) to control which Claude model each agent uses, balancing quality vs token spend. New files: - get-shit-done/references/model-profiles.md - profile lookup table - commands/gsd/set-profile.md - runtime profile switching Updated orchestrators to resolve model before spawning agents: - new-project, new-milestone (7 calls each) - plan-phase (4 calls) - execute-phase (3 calls) - audit-milestone, debug, research-phase, quick (1-2 calls each) - workflows: execute-phase, execute-plan, verify-work, map-codebase Profile stored in .planning/config.json as model_profile. Default: "balanced" (Opus for planning, Sonnet for execution). Closes #160 Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
This commit is contained in:
Lex Christopherson
20
README.md
20
README.md
@@ -485,7 +485,8 @@ You're never locked in. The system adapts.
|
||||
|
||||
| Command | What it does |
|
||||
|---------|--------------|
|
||||
| `/gsd:settings` | Toggle workflow agents (research, plan checker, verifier) |
|
||||
| `/gsd:settings` | Configure model profile and workflow agents |
|
||||
| `/gsd:set-profile <profile>` | Switch model profile (quality/balanced/budget) |
|
||||
| `/gsd:add-todo [desc]` | Capture idea for later |
|
||||
| `/gsd:check-todos` | List pending todos |
|
||||
| `/gsd:debug [desc]` | Systematic debugging with persistent state |
|
||||
@@ -506,6 +507,23 @@ GSD stores project settings in `.planning/config.json`. Configure during `/gsd:n
|
||||
| `mode` | `yolo`, `interactive` | `interactive` | Auto-approve vs confirm at each step |
|
||||
| `depth` | `quick`, `standard`, `comprehensive` | `standard` | Planning thoroughness (phases × plans) |
|
||||
|
||||
### Model Profiles
|
||||
|
||||
Control which Claude model each agent uses. Balance quality vs token spend.
|
||||
|
||||
| Profile | Planning | Execution | Verification |
|
||||
|---------|----------|-----------|--------------|
|
||||
| `quality` | Opus | Opus | Sonnet |
|
||||
| `balanced` (default) | Opus | Sonnet | Sonnet |
|
||||
| `budget` | Sonnet | Sonnet | Haiku |
|
||||
|
||||
Switch profiles:
|
||||
```
|
||||
/gsd:set-profile budget
|
||||
```
|
||||
|
||||
Or configure via `/gsd:settings`.
|
||||
|
||||
### Workflow Agents
|
||||
|
||||
These spawn additional agents during planning/execution. They improve quality but add tokens and time.
|
||||
|
||||
@@ -39,6 +39,24 @@ Glob: .planning/phases/*/*-VERIFICATION.md
|
||||
|
||||
<process>
|
||||
|
||||
## 0. Resolve Model Profile
|
||||
|
||||
Read model profile for agent spawning:
|
||||
|
||||
```bash
|
||||
MODEL_PROFILE=$(cat .planning/config.json 2>/dev/null | grep -o '"model_profile"[[:space:]]*:[[:space:]]*"[^"]*"' | grep -o '"[^"]*"$' | tr -d '"' || echo "balanced")
|
||||
```
|
||||
|
||||
Default to "balanced" if not set.
|
||||
|
||||
**Model lookup table:**
|
||||
|
||||
| Agent | quality | balanced | budget |
|
||||
|-------|---------|----------|--------|
|
||||
| gsd-integration-checker | sonnet | sonnet | haiku |
|
||||
|
||||
Store resolved model for use in Task call below.
|
||||
|
||||
## 1. Determine Milestone Scope
|
||||
|
||||
```bash
|
||||
@@ -83,7 +101,8 @@ Phase exports: {from SUMMARYs}
|
||||
API routes: {routes created}
|
||||
|
||||
Verify cross-phase wiring and E2E user flows.",
|
||||
subagent_type="gsd-integration-checker"
|
||||
subagent_type="gsd-integration-checker",
|
||||
model="{integration_checker_model}"
|
||||
)
|
||||
```
|
||||
|
||||
|
||||
@@ -28,6 +28,24 @@ ls .planning/debug/*.md 2>/dev/null | grep -v resolved | head -5
|
||||
|
||||
<process>
|
||||
|
||||
## 0. Resolve Model Profile
|
||||
|
||||
Read model profile for agent spawning:
|
||||
|
||||
```bash
|
||||
MODEL_PROFILE=$(cat .planning/config.json 2>/dev/null | grep -o '"model_profile"[[:space:]]*:[[:space:]]*"[^"]*"' | grep -o '"[^"]*"$' | tr -d '"' || echo "balanced")
|
||||
```
|
||||
|
||||
Default to "balanced" if not set.
|
||||
|
||||
**Model lookup table:**
|
||||
|
||||
| Agent | quality | balanced | budget |
|
||||
|-------|---------|----------|--------|
|
||||
| gsd-debugger | opus | sonnet | sonnet |
|
||||
|
||||
Store resolved model for use in Task calls below.
|
||||
|
||||
## 1. Check Active Sessions
|
||||
|
||||
If active sessions exist AND no $ARGUMENTS:
|
||||
@@ -82,6 +100,7 @@ Create: .planning/debug/{slug}.md
|
||||
Task(
|
||||
prompt=filled_prompt,
|
||||
subagent_type="gsd-debugger",
|
||||
model="{debugger_model}",
|
||||
description="Debug {slug}"
|
||||
)
|
||||
```
|
||||
@@ -134,6 +153,7 @@ goal: find_and_fix
|
||||
Task(
|
||||
prompt=continuation_prompt,
|
||||
subagent_type="gsd-debugger",
|
||||
model="{debugger_model}",
|
||||
description="Continue debug {slug}"
|
||||
)
|
||||
```
|
||||
|
||||
@@ -38,6 +38,24 @@ Phase: $ARGUMENTS
|
||||
</context>
|
||||
|
||||
<process>
|
||||
0. **Resolve Model Profile**
|
||||
|
||||
Read model profile for agent spawning:
|
||||
```bash
|
||||
MODEL_PROFILE=$(cat .planning/config.json 2>/dev/null | grep -o '"model_profile"[[:space:]]*:[[:space:]]*"[^"]*"' | grep -o '"[^"]*"$' | tr -d '"' || echo "balanced")
|
||||
```
|
||||
|
||||
Default to "balanced" if not set.
|
||||
|
||||
**Model lookup table:**
|
||||
|
||||
| Agent | quality | balanced | budget |
|
||||
|-------|---------|----------|--------|
|
||||
| gsd-executor | opus | sonnet | sonnet |
|
||||
| gsd-verifier | sonnet | sonnet | haiku |
|
||||
|
||||
Store resolved models for use in Task calls below.
|
||||
|
||||
1. **Validate phase exists**
|
||||
- Find phase directory matching argument
|
||||
- Count PLAN.md files
|
||||
@@ -248,9 +266,9 @@ STATE_CONTENT=$(cat .planning/STATE.md)
|
||||
Spawn all plans in a wave with a single message containing multiple Task calls, with inlined content:
|
||||
|
||||
```
|
||||
Task(prompt="Execute plan at {plan_01_path}\n\nPlan:\n{plan_01_content}\n\nProject state:\n{state_content}", subagent_type="gsd-executor")
|
||||
Task(prompt="Execute plan at {plan_02_path}\n\nPlan:\n{plan_02_content}\n\nProject state:\n{state_content}", subagent_type="gsd-executor")
|
||||
Task(prompt="Execute plan at {plan_03_path}\n\nPlan:\n{plan_03_content}\n\nProject state:\n{state_content}", subagent_type="gsd-executor")
|
||||
Task(prompt="Execute plan at {plan_01_path}\n\nPlan:\n{plan_01_content}\n\nProject state:\n{state_content}", subagent_type="gsd-executor", model="{executor_model}")
|
||||
Task(prompt="Execute plan at {plan_02_path}\n\nPlan:\n{plan_02_content}\n\nProject state:\n{state_content}", subagent_type="gsd-executor", model="{executor_model}")
|
||||
Task(prompt="Execute plan at {plan_03_path}\n\nPlan:\n{plan_03_content}\n\nProject state:\n{state_content}", subagent_type="gsd-executor", model="{executor_model}")
|
||||
```
|
||||
|
||||
All three run in parallel. Task tool blocks until all complete.
|
||||
|
||||
@@ -122,6 +122,26 @@ git add .planning/PROJECT.md .planning/STATE.md
|
||||
git commit -m "docs: start milestone v[X.Y] [Name]"
|
||||
```
|
||||
|
||||
## Phase 6.5: Resolve Model Profile
|
||||
|
||||
Read model profile for agent spawning:
|
||||
|
||||
```bash
|
||||
MODEL_PROFILE=$(cat .planning/config.json 2>/dev/null | grep -o '"model_profile"[[:space:]]*:[[:space:]]*"[^"]*"' | grep -o '"[^"]*"$' | tr -d '"' || echo "balanced")
|
||||
```
|
||||
|
||||
Default to "balanced" if not set.
|
||||
|
||||
**Model lookup table:**
|
||||
|
||||
| Agent | quality | balanced | budget |
|
||||
|-------|---------|----------|--------|
|
||||
| gsd-project-researcher | opus | sonnet | haiku |
|
||||
| gsd-research-synthesizer | sonnet | sonnet | haiku |
|
||||
| gsd-roadmapper | opus | sonnet | sonnet |
|
||||
|
||||
Store resolved models for use in Task calls below.
|
||||
|
||||
## Phase 7: Research Decision
|
||||
|
||||
Use AskUserQuestion:
|
||||
@@ -198,7 +218,7 @@ Your STACK.md feeds into roadmap creation. Be prescriptive:
|
||||
Write to: .planning/research/STACK.md
|
||||
Use template: ~/.claude/get-shit-done/templates/research-project/STACK.md
|
||||
</output>
|
||||
", subagent_type="gsd-project-researcher", description="Stack research")
|
||||
", subagent_type="gsd-project-researcher", model="{researcher_model}", description="Stack research")
|
||||
|
||||
Task(prompt="
|
||||
<research_type>
|
||||
@@ -239,7 +259,7 @@ Your FEATURES.md feeds into requirements definition. Categorize clearly:
|
||||
Write to: .planning/research/FEATURES.md
|
||||
Use template: ~/.claude/get-shit-done/templates/research-project/FEATURES.md
|
||||
</output>
|
||||
", subagent_type="gsd-project-researcher", description="Features research")
|
||||
", subagent_type="gsd-project-researcher", model="{researcher_model}", description="Features research")
|
||||
|
||||
Task(prompt="
|
||||
<research_type>
|
||||
@@ -281,7 +301,7 @@ Your ARCHITECTURE.md informs phase structure in roadmap. Include:
|
||||
Write to: .planning/research/ARCHITECTURE.md
|
||||
Use template: ~/.claude/get-shit-done/templates/research-project/ARCHITECTURE.md
|
||||
</output>
|
||||
", subagent_type="gsd-project-researcher", description="Architecture research")
|
||||
", subagent_type="gsd-project-researcher", model="{researcher_model}", description="Architecture research")
|
||||
|
||||
Task(prompt="
|
||||
<research_type>
|
||||
@@ -319,7 +339,7 @@ Your PITFALLS.md prevents mistakes in roadmap/planning. For each pitfall:
|
||||
Write to: .planning/research/PITFALLS.md
|
||||
Use template: ~/.claude/get-shit-done/templates/research-project/PITFALLS.md
|
||||
</output>
|
||||
", subagent_type="gsd-project-researcher", description="Pitfalls research")
|
||||
", subagent_type="gsd-project-researcher", model="{researcher_model}", description="Pitfalls research")
|
||||
```
|
||||
|
||||
After all 4 agents complete, spawn synthesizer to create SUMMARY.md:
|
||||
@@ -343,7 +363,7 @@ Write to: .planning/research/SUMMARY.md
|
||||
Use template: ~/.claude/get-shit-done/templates/research-project/SUMMARY.md
|
||||
Commit after writing.
|
||||
</output>
|
||||
", subagent_type="gsd-research-synthesizer", description="Synthesize research")
|
||||
", subagent_type="gsd-research-synthesizer", model="{synthesizer_model}", description="Synthesize research")
|
||||
```
|
||||
|
||||
Display research complete banner and key findings:
|
||||
@@ -548,7 +568,7 @@ Create roadmap for milestone v[X.Y]:
|
||||
|
||||
Write files first, then return. This ensures artifacts persist even if context is lost.
|
||||
</instructions>
|
||||
", subagent_type="gsd-roadmapper", description="Create roadmap")
|
||||
", subagent_type="gsd-roadmapper", model="{roadmapper_model}", description="Create roadmap")
|
||||
```
|
||||
|
||||
**Handle roadmapper return:**
|
||||
@@ -615,7 +635,7 @@ Use AskUserQuestion:
|
||||
Update the roadmap based on feedback. Edit files in place.
|
||||
Return ROADMAP REVISED with changes made.
|
||||
</revision>
|
||||
", subagent_type="gsd-roadmapper", description="Revise roadmap")
|
||||
", subagent_type="gsd-roadmapper", model="{roadmapper_model}", description="Revise roadmap")
|
||||
```
|
||||
- Present revised roadmap
|
||||
- Loop until user approves
|
||||
|
||||
@@ -373,6 +373,26 @@ EOF
|
||||
|
||||
**Note:** Run `/gsd:settings` anytime to update these preferences.
|
||||
|
||||
## Phase 5.5: Resolve Model Profile
|
||||
|
||||
Read model profile for agent spawning:
|
||||
|
||||
```bash
|
||||
MODEL_PROFILE=$(cat .planning/config.json 2>/dev/null | grep -o '"model_profile"[[:space:]]*:[[:space:]]*"[^"]*"' | grep -o '"[^"]*"$' | tr -d '"' || echo "balanced")
|
||||
```
|
||||
|
||||
Default to "balanced" if not set.
|
||||
|
||||
**Model lookup table:**
|
||||
|
||||
| Agent | quality | balanced | budget |
|
||||
|-------|---------|----------|--------|
|
||||
| gsd-project-researcher | opus | sonnet | haiku |
|
||||
| gsd-research-synthesizer | sonnet | sonnet | haiku |
|
||||
| gsd-roadmapper | opus | sonnet | sonnet |
|
||||
|
||||
Store resolved models for use in Task calls below.
|
||||
|
||||
## Phase 6: Research Decision
|
||||
|
||||
Use AskUserQuestion:
|
||||
@@ -453,7 +473,7 @@ Your STACK.md feeds into roadmap creation. Be prescriptive:
|
||||
Write to: .planning/research/STACK.md
|
||||
Use template: ~/.claude/get-shit-done/templates/research-project/STACK.md
|
||||
</output>
|
||||
", subagent_type="gsd-project-researcher", description="Stack research")
|
||||
", subagent_type="gsd-project-researcher", model="{researcher_model}", description="Stack research")
|
||||
|
||||
Task(prompt="
|
||||
<research_type>
|
||||
@@ -492,7 +512,7 @@ Your FEATURES.md feeds into requirements definition. Categorize clearly:
|
||||
Write to: .planning/research/FEATURES.md
|
||||
Use template: ~/.claude/get-shit-done/templates/research-project/FEATURES.md
|
||||
</output>
|
||||
", subagent_type="gsd-project-researcher", description="Features research")
|
||||
", subagent_type="gsd-project-researcher", model="{researcher_model}", description="Features research")
|
||||
|
||||
Task(prompt="
|
||||
<research_type>
|
||||
@@ -531,7 +551,7 @@ Your ARCHITECTURE.md informs phase structure in roadmap. Include:
|
||||
Write to: .planning/research/ARCHITECTURE.md
|
||||
Use template: ~/.claude/get-shit-done/templates/research-project/ARCHITECTURE.md
|
||||
</output>
|
||||
", subagent_type="gsd-project-researcher", description="Architecture research")
|
||||
", subagent_type="gsd-project-researcher", model="{researcher_model}", description="Architecture research")
|
||||
|
||||
Task(prompt="
|
||||
<research_type>
|
||||
@@ -570,7 +590,7 @@ Your PITFALLS.md prevents mistakes in roadmap/planning. For each pitfall:
|
||||
Write to: .planning/research/PITFALLS.md
|
||||
Use template: ~/.claude/get-shit-done/templates/research-project/PITFALLS.md
|
||||
</output>
|
||||
", subagent_type="gsd-project-researcher", description="Pitfalls research")
|
||||
", subagent_type="gsd-project-researcher", model="{researcher_model}", description="Pitfalls research")
|
||||
```
|
||||
|
||||
After all 4 agents complete, spawn synthesizer to create SUMMARY.md:
|
||||
@@ -594,7 +614,7 @@ Write to: .planning/research/SUMMARY.md
|
||||
Use template: ~/.claude/get-shit-done/templates/research-project/SUMMARY.md
|
||||
Commit after writing.
|
||||
</output>
|
||||
", subagent_type="gsd-research-synthesizer", description="Synthesize research")
|
||||
", subagent_type="gsd-research-synthesizer", model="{synthesizer_model}", description="Synthesize research")
|
||||
```
|
||||
|
||||
Display research complete banner and key findings:
|
||||
@@ -799,7 +819,7 @@ Create roadmap:
|
||||
|
||||
Write files first, then return. This ensures artifacts persist even if context is lost.
|
||||
</instructions>
|
||||
", subagent_type="gsd-roadmapper", description="Create roadmap")
|
||||
", subagent_type="gsd-roadmapper", model="{roadmapper_model}", description="Create roadmap")
|
||||
```
|
||||
|
||||
**Handle roadmapper return:**
|
||||
@@ -875,7 +895,7 @@ Use AskUserQuestion:
|
||||
Update the roadmap based on feedback. Edit files in place.
|
||||
Return ROADMAP REVISED with changes made.
|
||||
</revision>
|
||||
", subagent_type="gsd-roadmapper", description="Revise roadmap")
|
||||
", subagent_type="gsd-roadmapper", model="{roadmapper_model}", description="Revise roadmap")
|
||||
```
|
||||
- Present revised roadmap
|
||||
- Loop until user approves
|
||||
|
||||
@@ -42,7 +42,7 @@ Normalize phase input in step 2 before any directory lookups.
|
||||
|
||||
<process>
|
||||
|
||||
## 1. Validate Environment
|
||||
## 1. Validate Environment and Resolve Model Profile
|
||||
|
||||
```bash
|
||||
ls .planning/ 2>/dev/null
|
||||
@@ -50,6 +50,24 @@ ls .planning/ 2>/dev/null
|
||||
|
||||
**If not found:** Error - user should run `/gsd:new-project` first.
|
||||
|
||||
**Resolve model profile for agent spawning:**
|
||||
|
||||
```bash
|
||||
MODEL_PROFILE=$(cat .planning/config.json 2>/dev/null | grep -o '"model_profile"[[:space:]]*:[[:space:]]*"[^"]*"' | grep -o '"[^"]*"$' | tr -d '"' || echo "balanced")
|
||||
```
|
||||
|
||||
Default to "balanced" if not set.
|
||||
|
||||
**Model lookup table:**
|
||||
|
||||
| Agent | quality | balanced | budget |
|
||||
|-------|---------|----------|--------|
|
||||
| gsd-phase-researcher | opus | sonnet | haiku |
|
||||
| gsd-planner | opus | opus | sonnet |
|
||||
| gsd-plan-checker | sonnet | sonnet | haiku |
|
||||
|
||||
Store resolved models for use in Task calls below.
|
||||
|
||||
## 2. Parse and Normalize Arguments
|
||||
|
||||
Extract from $ARGUMENTS:
|
||||
@@ -190,6 +208,7 @@ Write research findings to: {phase_dir}/{phase}-RESEARCH.md
|
||||
Task(
|
||||
prompt=research_prompt,
|
||||
subagent_type="gsd-phase-researcher",
|
||||
model="{researcher_model}",
|
||||
description="Research Phase {phase}"
|
||||
)
|
||||
```
|
||||
@@ -298,6 +317,7 @@ Before returning PLANNING COMPLETE:
|
||||
Task(
|
||||
prompt=filled_prompt,
|
||||
subagent_type="gsd-planner",
|
||||
model="{planner_model}",
|
||||
description="Plan Phase {phase}"
|
||||
)
|
||||
```
|
||||
@@ -369,6 +389,7 @@ Return one of:
|
||||
Task(
|
||||
prompt=checker_prompt,
|
||||
subagent_type="gsd-plan-checker",
|
||||
model="{checker_model}",
|
||||
description="Verify Phase {phase} plans"
|
||||
)
|
||||
```
|
||||
@@ -426,6 +447,7 @@ Return what changed.
|
||||
Task(
|
||||
prompt=revision_prompt,
|
||||
subagent_type="gsd-planner",
|
||||
model="{planner_model}",
|
||||
description="Revise Phase {phase} plans"
|
||||
)
|
||||
```
|
||||
|
||||
@@ -52,6 +52,7 @@ If missing both ROADMAP.md and PROJECT.md: suggest `/gsd:new-project`.
|
||||
- Read `.planning/STATE.md` for living memory (position, decisions, issues)
|
||||
- Read `.planning/ROADMAP.md` for phase structure and objectives
|
||||
- Read `.planning/PROJECT.md` for current state (What This Is, Core Value, Requirements)
|
||||
- Read `.planning/config.json` for settings (model_profile, workflow toggles)
|
||||
</step>
|
||||
|
||||
<step name="recent">
|
||||
@@ -80,6 +81,7 @@ If missing both ROADMAP.md and PROJECT.md: suggest `/gsd:new-project`.
|
||||
# [Project Name]
|
||||
|
||||
**Progress:** [████████░░] 8/10 plans complete
|
||||
**Profile:** [quality/balanced/budget]
|
||||
|
||||
## Recent Work
|
||||
- [Phase X, Plan Y]: [what was accomplished - 1 line]
|
||||
|
||||
@@ -34,6 +34,27 @@ Orchestration is inline - no separate workflow file. Quick mode is deliberately
|
||||
</context>
|
||||
|
||||
<process>
|
||||
**Step 0: Resolve Model Profile**
|
||||
|
||||
Read model profile for agent spawning:
|
||||
|
||||
```bash
|
||||
MODEL_PROFILE=$(cat .planning/config.json 2>/dev/null | grep -o '"model_profile"[[:space:]]*:[[:space:]]*"[^"]*"' | grep -o '"[^"]*"$' | tr -d '"' || echo "balanced")
|
||||
```
|
||||
|
||||
Default to "balanced" if not set.
|
||||
|
||||
**Model lookup table:**
|
||||
|
||||
| Agent | quality | balanced | budget |
|
||||
|-------|---------|----------|--------|
|
||||
| gsd-planner | opus | opus | sonnet |
|
||||
| gsd-executor | opus | sonnet | sonnet |
|
||||
|
||||
Store resolved models for use in Task calls below.
|
||||
|
||||
---
|
||||
|
||||
**Step 1: Pre-flight validation**
|
||||
|
||||
Check that an active GSD project exists:
|
||||
@@ -145,6 +166,7 @@ Return: ## PLANNING COMPLETE with plan path
|
||||
</output>
|
||||
",
|
||||
subagent_type="gsd-planner",
|
||||
model="{planner_model}",
|
||||
description="Quick plan: ${DESCRIPTION}"
|
||||
)
|
||||
```
|
||||
@@ -178,6 +200,7 @@ Project state: @.planning/STATE.md
|
||||
</constraints>
|
||||
",
|
||||
subagent_type="gsd-executor",
|
||||
model="{executor_model}",
|
||||
description="Execute: ${DESCRIPTION}"
|
||||
)
|
||||
```
|
||||
|
||||
@@ -31,6 +31,24 @@ Normalize phase input in step 1 before any directory lookups.
|
||||
|
||||
<process>
|
||||
|
||||
## 0. Resolve Model Profile
|
||||
|
||||
Read model profile for agent spawning:
|
||||
|
||||
```bash
|
||||
MODEL_PROFILE=$(cat .planning/config.json 2>/dev/null | grep -o '"model_profile"[[:space:]]*:[[:space:]]*"[^"]*"' | grep -o '"[^"]*"$' | tr -d '"' || echo "balanced")
|
||||
```
|
||||
|
||||
Default to "balanced" if not set.
|
||||
|
||||
**Model lookup table:**
|
||||
|
||||
| Agent | quality | balanced | budget |
|
||||
|-------|---------|----------|--------|
|
||||
| gsd-phase-researcher | opus | sonnet | haiku |
|
||||
|
||||
Store resolved model for use in Task calls below.
|
||||
|
||||
## 1. Normalize and Validate Phase
|
||||
|
||||
```bash
|
||||
@@ -132,6 +150,7 @@ Write to: .planning/phases/${PHASE}-{slug}/${PHASE}-RESEARCH.md
|
||||
Task(
|
||||
prompt=filled_prompt,
|
||||
subagent_type="gsd-phase-researcher",
|
||||
model="{researcher_model}",
|
||||
description="Research Phase {phase}"
|
||||
)
|
||||
```
|
||||
@@ -165,6 +184,7 @@ Research file: @.planning/phases/${PHASE}-{slug}/${PHASE}-RESEARCH.md
|
||||
Task(
|
||||
prompt=continuation_prompt,
|
||||
subagent_type="gsd-phase-researcher",
|
||||
model="{researcher_model}",
|
||||
description="Continue research Phase {phase}"
|
||||
)
|
||||
```
|
||||
|
||||
106
commands/gsd/set-profile.md
Normal file
106
commands/gsd/set-profile.md
Normal file
@@ -0,0 +1,106 @@
|
||||
---
|
||||
name: set-profile
|
||||
description: Switch model profile for GSD agents (quality/balanced/budget)
|
||||
arguments:
|
||||
- name: profile
|
||||
description: "Profile name: quality, balanced, or budget"
|
||||
required: true
|
||||
---
|
||||
|
||||
<objective>
|
||||
Switch the model profile used by GSD agents. This controls which Claude model each agent uses, balancing quality vs token spend.
|
||||
</objective>
|
||||
|
||||
<profiles>
|
||||
| Profile | Description |
|
||||
|---------|-------------|
|
||||
| **quality** | Opus everywhere except read-only verification |
|
||||
| **balanced** | Opus for planning, Sonnet for execution/verification (default) |
|
||||
| **budget** | Sonnet for writing, Haiku for research/verification |
|
||||
</profiles>
|
||||
|
||||
<process>
|
||||
|
||||
## 1. Validate argument
|
||||
|
||||
```
|
||||
if $ARGUMENTS.profile not in ["quality", "balanced", "budget"]:
|
||||
Error: Invalid profile "$ARGUMENTS.profile"
|
||||
Valid profiles: quality, balanced, budget
|
||||
STOP
|
||||
```
|
||||
|
||||
## 2. Check for project
|
||||
|
||||
```bash
|
||||
ls .planning/config.json 2>/dev/null
|
||||
```
|
||||
|
||||
If no `.planning/` directory:
|
||||
```
|
||||
Error: No GSD project found.
|
||||
Run /gsd:new-project first to initialize a project.
|
||||
```
|
||||
|
||||
## 3. Update config.json
|
||||
|
||||
Read current config:
|
||||
```bash
|
||||
cat .planning/config.json
|
||||
```
|
||||
|
||||
Update `model_profile` field (or add if missing):
|
||||
```json
|
||||
{
|
||||
"model_profile": "$ARGUMENTS.profile"
|
||||
}
|
||||
```
|
||||
|
||||
Write updated config back to `.planning/config.json`.
|
||||
|
||||
## 4. Confirm
|
||||
|
||||
```
|
||||
✓ Model profile set to: $ARGUMENTS.profile
|
||||
|
||||
Agents will now use:
|
||||
[Show table from model-profiles.md for selected profile]
|
||||
|
||||
Next spawned agents will use the new profile.
|
||||
```
|
||||
|
||||
</process>
|
||||
|
||||
<examples>
|
||||
|
||||
**Switch to budget mode:**
|
||||
```
|
||||
/gsd:set-profile budget
|
||||
|
||||
✓ Model profile set to: budget
|
||||
|
||||
Agents will now use:
|
||||
| Agent | Model |
|
||||
|-------|-------|
|
||||
| gsd-planner | sonnet |
|
||||
| gsd-executor | sonnet |
|
||||
| gsd-verifier | haiku |
|
||||
| ... | ... |
|
||||
```
|
||||
|
||||
**Switch to quality mode:**
|
||||
```
|
||||
/gsd:set-profile quality
|
||||
|
||||
✓ Model profile set to: quality
|
||||
|
||||
Agents will now use:
|
||||
| Agent | Model |
|
||||
|-------|-------|
|
||||
| gsd-planner | opus |
|
||||
| gsd-executor | opus |
|
||||
| gsd-verifier | sonnet |
|
||||
| ... | ... |
|
||||
```
|
||||
|
||||
</examples>
|
||||
@@ -1,6 +1,6 @@
|
||||
---
|
||||
name: gsd:settings
|
||||
description: Configure GSD workflow toggles (research, plan checker, verifier)
|
||||
description: Configure GSD workflow toggles and model profile
|
||||
allowed-tools:
|
||||
- Read
|
||||
- Write
|
||||
@@ -8,9 +8,9 @@ allowed-tools:
|
||||
---
|
||||
|
||||
<objective>
|
||||
Allow users to toggle workflow agents on/off via interactive settings.
|
||||
Allow users to toggle workflow agents on/off and select model profile via interactive settings.
|
||||
|
||||
Updates `.planning/config.json` with workflow preferences that control whether research, plan checking, and verification agents are spawned.
|
||||
Updates `.planning/config.json` with workflow preferences and model profile selection.
|
||||
</objective>
|
||||
|
||||
<process>
|
||||
@@ -33,6 +33,7 @@ Parse current values (default to `true` if not present):
|
||||
- `workflow.research` — spawn researcher during plan-phase
|
||||
- `workflow.plan_check` — spawn plan checker during plan-phase
|
||||
- `workflow.verifier` — spawn verifier during execute-phase
|
||||
- `model_profile` — which model each agent uses (default: `balanced`)
|
||||
|
||||
## 3. Present Settings
|
||||
|
||||
@@ -40,6 +41,16 @@ Use AskUserQuestion with current values shown:
|
||||
|
||||
```
|
||||
AskUserQuestion([
|
||||
{
|
||||
question: "Which model profile for agents?",
|
||||
header: "Model",
|
||||
multiSelect: false,
|
||||
options: [
|
||||
{ label: "Quality", description: "Opus everywhere except verification (highest cost)" },
|
||||
{ label: "Balanced (Recommended)", description: "Opus for planning, Sonnet for execution/verification" },
|
||||
{ label: "Budget", description: "Sonnet for writing, Haiku for research/verification (lowest cost)" }
|
||||
]
|
||||
},
|
||||
{
|
||||
question: "Spawn Plan Researcher? (researches domain before planning)",
|
||||
header: "Research",
|
||||
@@ -74,11 +85,12 @@ AskUserQuestion([
|
||||
|
||||
## 4. Update Config
|
||||
|
||||
Merge new workflow settings into existing config.json:
|
||||
Merge new settings into existing config.json:
|
||||
|
||||
```json
|
||||
{
|
||||
...existing_config,
|
||||
"model_profile": "quality" | "balanced" | "budget",
|
||||
"workflow": {
|
||||
"research": true/false,
|
||||
"plan_check": true/false,
|
||||
@@ -100,23 +112,25 @@ Display:
|
||||
|
||||
| Setting | Value |
|
||||
|----------------------|-------|
|
||||
| Model Profile | {quality/balanced/budget} |
|
||||
| Plan Researcher | {On/Off} |
|
||||
| Plan Checker | {On/Off} |
|
||||
| Execution Verifier | {On/Off} |
|
||||
|
||||
These settings apply to future /gsd:plan-phase and /gsd:execute-phase runs.
|
||||
|
||||
Override per-invocation with flags:
|
||||
- /gsd:plan-phase --research (force research)
|
||||
- /gsd:plan-phase --skip-research (skip research)
|
||||
- /gsd:plan-phase --skip-verify (skip plan check)
|
||||
Quick commands:
|
||||
- /gsd:set-profile <profile> — switch model profile
|
||||
- /gsd:plan-phase --research — force research
|
||||
- /gsd:plan-phase --skip-research — skip research
|
||||
- /gsd:plan-phase --skip-verify — skip plan check
|
||||
```
|
||||
|
||||
</process>
|
||||
|
||||
<success_criteria>
|
||||
- [ ] Current config read
|
||||
- [ ] User presented with 3 toggle questions
|
||||
- [ ] Config updated with workflow section
|
||||
- [ ] User presented with 4 settings (profile + 3 toggles)
|
||||
- [ ] Config updated with model_profile and workflow section
|
||||
- [ ] Changes confirmed to user
|
||||
</success_criteria>
|
||||
|
||||
73
get-shit-done/references/model-profiles.md
Normal file
73
get-shit-done/references/model-profiles.md
Normal file
@@ -0,0 +1,73 @@
|
||||
# Model Profiles
|
||||
|
||||
Model profiles control which Claude model each GSD agent uses. This allows balancing quality vs token spend.
|
||||
|
||||
## Profile Definitions
|
||||
|
||||
| Agent | `quality` | `balanced` | `budget` |
|
||||
|-------|-----------|------------|----------|
|
||||
| gsd-planner | opus | opus | sonnet |
|
||||
| gsd-roadmapper | opus | sonnet | sonnet |
|
||||
| gsd-executor | opus | sonnet | sonnet |
|
||||
| gsd-phase-researcher | opus | sonnet | haiku |
|
||||
| gsd-project-researcher | opus | sonnet | haiku |
|
||||
| gsd-research-synthesizer | sonnet | sonnet | haiku |
|
||||
| gsd-debugger | opus | sonnet | sonnet |
|
||||
| gsd-codebase-mapper | sonnet | haiku | haiku |
|
||||
| gsd-verifier | sonnet | sonnet | haiku |
|
||||
| gsd-plan-checker | sonnet | sonnet | haiku |
|
||||
| gsd-integration-checker | sonnet | sonnet | haiku |
|
||||
|
||||
## Profile Philosophy
|
||||
|
||||
**quality** - Maximum reasoning power
|
||||
- Opus for all decision-making agents
|
||||
- Sonnet for read-only verification
|
||||
- Use when: quota available, critical architecture work
|
||||
|
||||
**balanced** (default) - Smart allocation
|
||||
- Opus only for planning (where architecture decisions happen)
|
||||
- Sonnet for execution and research (follows explicit instructions)
|
||||
- Sonnet for verification (needs reasoning, not just pattern matching)
|
||||
- Use when: normal development, good balance of quality and cost
|
||||
|
||||
**budget** - Minimal Opus usage
|
||||
- Sonnet for anything that writes code
|
||||
- Haiku for research and verification
|
||||
- Use when: conserving quota, high-volume work, less critical phases
|
||||
|
||||
## Resolution Logic
|
||||
|
||||
Orchestrators resolve model before spawning:
|
||||
|
||||
```
|
||||
1. Read .planning/config.json
|
||||
2. Get model_profile (default: "balanced")
|
||||
3. Look up agent in table above
|
||||
4. Pass model parameter to Task call
|
||||
```
|
||||
|
||||
## Switching Profiles
|
||||
|
||||
Runtime: `/gsd:set-profile <profile>`
|
||||
|
||||
Per-project default: Set in `.planning/config.json`:
|
||||
```json
|
||||
{
|
||||
"model_profile": "balanced"
|
||||
}
|
||||
```
|
||||
|
||||
## Design Rationale
|
||||
|
||||
**Why Opus for gsd-planner?**
|
||||
Planning involves architecture decisions, goal decomposition, and task design. This is where model quality has the highest impact.
|
||||
|
||||
**Why Sonnet for gsd-executor?**
|
||||
Executors follow explicit PLAN.md instructions. The plan already contains the reasoning; execution is implementation.
|
||||
|
||||
**Why Sonnet (not Haiku) for verifiers in balanced?**
|
||||
Verification requires goal-backward reasoning - checking if code *delivers* what the phase promised, not just pattern matching. Sonnet handles this well; Haiku may miss subtle gaps.
|
||||
|
||||
**Why Haiku for gsd-codebase-mapper?**
|
||||
Read-only exploration and pattern extraction. No reasoning required, just structured output from file contents.
|
||||
@@ -13,7 +13,27 @@ Read config.json for planning behavior settings.
|
||||
|
||||
<process>
|
||||
|
||||
<step name="load_project_state" priority="first">
|
||||
<step name="resolve_model_profile" priority="first">
|
||||
Read model profile for agent spawning:
|
||||
|
||||
```bash
|
||||
MODEL_PROFILE=$(cat .planning/config.json 2>/dev/null | grep -o '"model_profile"[[:space:]]*:[[:space:]]*"[^"]*"' | grep -o '"[^"]*"$' | tr -d '"' || echo "balanced")
|
||||
```
|
||||
|
||||
Default to "balanced" if not set.
|
||||
|
||||
**Model lookup table:**
|
||||
|
||||
| Agent | quality | balanced | budget |
|
||||
|-------|---------|----------|--------|
|
||||
| gsd-executor | opus | sonnet | sonnet |
|
||||
| gsd-verifier | sonnet | sonnet | haiku |
|
||||
| general-purpose | — | — | — |
|
||||
|
||||
Store resolved models for use in Task calls below.
|
||||
</step>
|
||||
|
||||
<step name="load_project_state">
|
||||
Before any operation, read project state:
|
||||
|
||||
```bash
|
||||
@@ -274,7 +294,7 @@ Plans with `autonomous: false` require user interaction.
|
||||
|
||||
1. **Spawn agent for checkpoint plan:**
|
||||
```
|
||||
Task(prompt="{subagent-task-prompt}", subagent_type="general-purpose")
|
||||
Task(prompt="{subagent-task-prompt}", subagent_type="gsd-executor", model="{executor_model}")
|
||||
```
|
||||
|
||||
2. **Agent runs until checkpoint:**
|
||||
@@ -313,7 +333,8 @@ Plans with `autonomous: false` require user interaction.
|
||||
```
|
||||
Task(
|
||||
prompt=filled_continuation_template,
|
||||
subagent_type="general-purpose"
|
||||
subagent_type="gsd-executor",
|
||||
model="{executor_model}"
|
||||
)
|
||||
```
|
||||
|
||||
@@ -389,7 +410,8 @@ Phase goal: {goal from ROADMAP.md}
|
||||
|
||||
Check must_haves against actual codebase. Create VERIFICATION.md.
|
||||
Verify what actually exists in the code.",
|
||||
subagent_type="gsd-verifier"
|
||||
subagent_type="gsd-verifier",
|
||||
model="{verifier_model}"
|
||||
)
|
||||
```
|
||||
|
||||
|
||||
@@ -11,7 +11,25 @@ Read config.json for planning behavior settings.
|
||||
|
||||
<process>
|
||||
|
||||
<step name="load_project_state" priority="first">
|
||||
<step name="resolve_model_profile" priority="first">
|
||||
Read model profile for agent spawning:
|
||||
|
||||
```bash
|
||||
MODEL_PROFILE=$(cat .planning/config.json 2>/dev/null | grep -o '"model_profile"[[:space:]]*:[[:space:]]*"[^"]*"' | grep -o '"[^"]*"$' | tr -d '"' || echo "balanced")
|
||||
```
|
||||
|
||||
Default to "balanced" if not set.
|
||||
|
||||
**Model lookup table:**
|
||||
|
||||
| Agent | quality | balanced | budget |
|
||||
|-------|---------|----------|--------|
|
||||
| gsd-executor | opus | sonnet | sonnet |
|
||||
|
||||
Store resolved model for use in Task calls below.
|
||||
</step>
|
||||
|
||||
<step name="load_project_state">
|
||||
Before any operation, read project state:
|
||||
|
||||
```bash
|
||||
@@ -219,7 +237,7 @@ No segmentation benefit - execute entirely in main
|
||||
```
|
||||
1. Run init_agent_tracking step first (see step below)
|
||||
|
||||
2. Use Task tool with subagent_type="gsd-executor":
|
||||
2. Use Task tool with subagent_type="gsd-executor" and model="{executor_model}":
|
||||
|
||||
Prompt: "Execute plan at .planning/phases/{phase}-{plan}-PLAN.md
|
||||
|
||||
@@ -371,7 +389,7 @@ For Pattern A (fully autonomous) and Pattern C (decision-dependent), skip this s
|
||||
|
||||
B. If routing = Subagent:
|
||||
```
|
||||
Spawn Task tool with subagent_type="gsd-executor":
|
||||
Spawn Task tool with subagent_type="gsd-executor" and model="{executor_model}":
|
||||
|
||||
Prompt: "Execute tasks [task numbers/names] from plan at [plan path].
|
||||
|
||||
|
||||
@@ -22,7 +22,25 @@ Documents are reference material for Claude when planning/executing. Always incl
|
||||
|
||||
<process>
|
||||
|
||||
<step name="check_existing" priority="first">
|
||||
<step name="resolve_model_profile" priority="first">
|
||||
Read model profile for agent spawning:
|
||||
|
||||
```bash
|
||||
MODEL_PROFILE=$(cat .planning/config.json 2>/dev/null | grep -o '"model_profile"[[:space:]]*:[[:space:]]*"[^"]*"' | grep -o '"[^"]*"$' | tr -d '"' || echo "balanced")
|
||||
```
|
||||
|
||||
Default to "balanced" if not set.
|
||||
|
||||
**Model lookup table:**
|
||||
|
||||
| Agent | quality | balanced | budget |
|
||||
|-------|---------|----------|--------|
|
||||
| gsd-codebase-mapper | sonnet | haiku | haiku |
|
||||
|
||||
Store resolved model for use in Task calls below.
|
||||
</step>
|
||||
|
||||
<step name="check_existing">
|
||||
Check if .planning/codebase/ already exists:
|
||||
|
||||
```bash
|
||||
@@ -73,7 +91,7 @@ Continue to spawn_agents.
|
||||
<step name="spawn_agents">
|
||||
Spawn 4 parallel gsd-codebase-mapper agents.
|
||||
|
||||
Use Task tool with `subagent_type="gsd-codebase-mapper"` and `run_in_background=true` for parallel execution.
|
||||
Use Task tool with `subagent_type="gsd-codebase-mapper"`, `model="{mapper_model}"`, and `run_in_background=true` for parallel execution.
|
||||
|
||||
**CRITICAL:** Use the dedicated `gsd-codebase-mapper` agent, NOT `Explore`. The mapper agent writes documents directly.
|
||||
|
||||
@@ -82,6 +100,7 @@ Use Task tool with `subagent_type="gsd-codebase-mapper"` and `run_in_background=
|
||||
Task tool parameters:
|
||||
```
|
||||
subagent_type: "gsd-codebase-mapper"
|
||||
model: "{mapper_model}"
|
||||
run_in_background: true
|
||||
description: "Map codebase tech stack"
|
||||
```
|
||||
@@ -104,6 +123,7 @@ Explore thoroughly. Write documents directly using templates. Return confirmatio
|
||||
Task tool parameters:
|
||||
```
|
||||
subagent_type: "gsd-codebase-mapper"
|
||||
model: "{mapper_model}"
|
||||
run_in_background: true
|
||||
description: "Map codebase architecture"
|
||||
```
|
||||
@@ -126,6 +146,7 @@ Explore thoroughly. Write documents directly using templates. Return confirmatio
|
||||
Task tool parameters:
|
||||
```
|
||||
subagent_type: "gsd-codebase-mapper"
|
||||
model: "{mapper_model}"
|
||||
run_in_background: true
|
||||
description: "Map codebase conventions"
|
||||
```
|
||||
@@ -148,6 +169,7 @@ Explore thoroughly. Write documents directly using templates. Return confirmatio
|
||||
Task tool parameters:
|
||||
```
|
||||
subagent_type: "gsd-codebase-mapper"
|
||||
model: "{mapper_model}"
|
||||
run_in_background: true
|
||||
description: "Map codebase concerns"
|
||||
```
|
||||
|
||||
@@ -20,6 +20,25 @@ No Pass/Fail buttons. No severity questions. Just: "Here's what should happen. D
|
||||
|
||||
<process>
|
||||
|
||||
<step name="resolve_model_profile" priority="first">
|
||||
Read model profile for agent spawning:
|
||||
|
||||
```bash
|
||||
MODEL_PROFILE=$(cat .planning/config.json 2>/dev/null | grep -o '"model_profile"[[:space:]]*:[[:space:]]*"[^"]*"' | grep -o '"[^"]*"$' | tr -d '"' || echo "balanced")
|
||||
```
|
||||
|
||||
Default to "balanced" if not set.
|
||||
|
||||
**Model lookup table:**
|
||||
|
||||
| Agent | quality | balanced | budget |
|
||||
|-------|---------|----------|--------|
|
||||
| gsd-planner | opus | opus | sonnet |
|
||||
| gsd-plan-checker | sonnet | sonnet | haiku |
|
||||
|
||||
Store resolved models for use in Task calls below.
|
||||
</step>
|
||||
|
||||
<step name="check_active_session">
|
||||
**First: Check for active UAT sessions**
|
||||
|
||||
@@ -389,6 +408,7 @@ Plans must be executable prompts.
|
||||
</downstream_consumer>
|
||||
""",
|
||||
subagent_type="gsd-planner",
|
||||
model="{planner_model}",
|
||||
description="Plan gap fixes for Phase {phase}"
|
||||
)
|
||||
```
|
||||
@@ -434,6 +454,7 @@ Return one of:
|
||||
</expected_output>
|
||||
""",
|
||||
subagent_type="gsd-plan-checker",
|
||||
model="{checker_model}",
|
||||
description="Verify Phase {phase} fix plans"
|
||||
)
|
||||
```
|
||||
@@ -474,6 +495,7 @@ Do NOT replan from scratch unless issues are fundamental.
|
||||
</instructions>
|
||||
""",
|
||||
subagent_type="gsd-planner",
|
||||
model="{planner_model}",
|
||||
description="Revise Phase {phase} plans"
|
||||
)
|
||||
```
|
||||
|
||||
Reference in New Issue
Block a user