Files
msd-core/get-shit-done/workflows/set-profile.md
TÂCHES d2623e0114 refactor: extract settings, add-todo, pause-work, set-profile to thin orchestrators
Completes Phase 1 thin orchestrator extraction.
Token savings: ~8,500 chars across 4 commands

Ported from: get-shit-done-v2@95b6214
2026-02-07 11:08:35 -06:00

1.7 KiB

Switch the model profile used by GSD agents. Controls which Claude model each agent uses, balancing quality vs token spend.

<required_reading> Read all files referenced by the invoking prompt's execution_context before starting. </required_reading>

Validate argument:
if $ARGUMENTS.profile not in ["quality", "balanced", "budget"]:
  Error: Invalid profile "$ARGUMENTS.profile"
  Valid profiles: quality, balanced, budget
  EXIT
```bash node ~/.claude/get-shit-done/bin/gsd-tools.js config-ensure-section ```

Creates .planning/config.json with defaults if missing.

Read current config: ```bash cat .planning/config.json ```

Update model_profile field:

{
  "model_profile": "$ARGUMENTS.profile"
}

Write updated config back to .planning/config.json.

Display confirmation with model table for selected profile:
✓ Model profile set to: $ARGUMENTS.profile

Agents will now use:

[Show table from MODEL_PROFILES in gsd-tools.js for selected profile]

Example:
| Agent | Model |
|-------|-------|
| gsd-planner | opus |
| gsd-executor | sonnet |
| gsd-verifier | haiku |
| ... | ... |

Next spawned agents will use the new profile.

Map profile names:

  • quality: use "quality" column from MODEL_PROFILES
  • balanced: use "balanced" column from MODEL_PROFILES
  • budget: use "budget" column from MODEL_PROFILES

<success_criteria>

  • Argument validated
  • Config file ensured
  • Config updated with new model_profile
  • Confirmation displayed with model table </success_criteria>