fix(#3936): dispatch quick research with researcher persona and model (#4158)

* test(#3936): quick research dispatch must use researcher persona/model

Regression tests for the planner-persona leak in quick's Step 4.75
research dispatch: init quick emits researcher_model, the workflow
resolves AGENT_SKILLS_RESEARCHER and parses researcher_model, and the
researcher dispatch uses the researcher persona/tier while planner and
executor dispatches keep their own.

* fix(#3936): dispatch quick research with researcher persona and model

cmdInitQuick now emits researcher_model (parity with the plan-phase
init), quick.md resolves AGENT_SKILLS_RESEARCHER and parses the new
field, and Step 4.75's dispatch swaps ${AGENT_SKILLS_PLANNER}/
{planner_model} for the researcher's own persona and tier. The #2517
model-omission list gains researcher_model.

Emitted-Drift-Ack-Growth: quick.md — researcher persona resolution line + researcher_model parse-list entry (#3936)

* chore(#3936): backfill changeset pr number

---------

Co-authored-by: sim <sim@local>

.changeset/silly-pandas-frolic.md

@@ -0,0 +1,5 @@
+---
+type: Fixed
+pr: 4158
+---
+**`/gsd-quick` research dispatch uses the researcher persona and model tier** — the quick flow's research step no longer injects the planner persona and planner model into `gsd-phase-researcher`; `init quick` now emits `researcher_model` and the workflow resolves `AGENT_SKILLS_RESEARCHER`, matching `/gsd-plan-phase`. (#3936)

gsd-core/workflows/quick.md

@@ -127,9 +127,10 @@ AGENT_SKILLS_PLANNER=$(gsd_run query agent-skills gsd-planner)
 AGENT_SKILLS_EXECUTOR=$(gsd_run query agent-skills gsd-executor)
 AGENT_SKILLS_CHECKER=$(gsd_run query agent-skills gsd-plan-checker)
 AGENT_SKILLS_VERIFIER=$(gsd_run query agent-skills gsd-verifier)
+AGENT_SKILLS_RESEARCHER=$(gsd_run query agent-skills gsd-phase-researcher)
 ```
 
-Parse JSON for: `planner_model`, `executor_model`, `checker_model`, `verifier_model`, `reviewer_model`, `commit_docs`, `branch_name`, `quick_id`, `slug`, `date`, `timestamp`, `quick_dir`, `task_dir`, `roadmap_exists`, `planning_exists`, `response_language`.
+Parse JSON for: `planner_model`, `executor_model`, `checker_model`, `verifier_model`, `reviewer_model`, `researcher_model`, `commit_docs`, `branch_name`, `quick_id`, `slug`, `date`, `timestamp`, `quick_dir`, `task_dir`, `roadmap_exists`, `planning_exists`, `response_language`.
 
 `init.quick` does not emit dedicated `state_path`/`project_path` fields, so derive them from the already-absolute `quick_dir` (#2376 — files handed to a spawned subagent must resolve regardless of that subagent's own cwd):
 ```bash

gsd-core/workflows/quick/steps/research-phase.md

@@ -13,7 +13,7 @@ Spawn a single focused researcher (not 4 parallel researchers like full phases
 
 <!-- #2517 model-omit-on-inherit -->
 
-> **Model omission (#2517).** Omit the `model` parameter entirely when the value it would carry (`planner_model`, `checker_model`, `executor_model`, `reviewer_model`, `verifier_model`) is `"inherit"` or empty. An empty value 404s on runtimes without native tier aliases — the default on non-Claude runtimes. Omitting it inherits the orchestrator's model. See @gsd-core/references/model-profile-resolution.md.
+> **Model omission (#2517).** Omit the `model` parameter entirely when the value it would carry (`planner_model`, `checker_model`, `executor_model`, `reviewer_model`, `verifier_model`, `researcher_model`) is `"inherit"` or empty. An empty value 404s on runtimes without native tier aliases — the default on non-Claude runtimes. Omitting it inherits the orchestrator's model. See @gsd-core/references/model-profile-resolution.md.
 
 <!-- #2508 runtime-aware-dispatch -->
 
@@ -35,7 +35,7 @@ Agent(
 ${DISCUSS_MODE ? '- ' + QUICK_DIR + '/' + quick_id + '-CONTEXT.md (User decisions — research should align with these. #3894: when workflow.research_before_questions is enabled research runs BEFORE discussion, so this file will not exist yet — read it only if present)' : ''}
 </required_reading>
 
-${AGENT_SKILLS_PLANNER}
+${AGENT_SKILLS_RESEARCHER}
 
 </research_context>
 
@@ -56,7 +56,7 @@ Return: ## RESEARCH COMPLETE with file path
 </output>
 ",
   subagent_type="gsd-phase-researcher",
-  model="{planner_model}",
+  model="{researcher_model}",
   description="Research: ${DESCRIPTION}"
 )
 ```

src/init.cts

@@ -1467,6 +1467,10 @@ function cmdInitQuick(
     executor_model: resolveModelInternal(cwd, 'gsd-executor'),
     checker_model: resolveModelInternal(cwd, 'gsd-plan-checker'),
     verifier_model: resolveModelInternal(cwd, 'gsd-verifier'),
+    // #3936: Step 4.75 dispatches gsd-phase-researcher; resolve its own tier
+    // (parity with cmdInitPlanPhase) so the research spawn stops pinning
+    // planner_model.
+    researcher_model: resolveModelInternal(cwd, 'gsd-phase-researcher'),
     // #2072: the quick review step spawns gsd-code-reviewer; resolve its own model
     // so model_overrides / models.verification apply (was reusing executor_model).
     reviewer_model: resolveModelInternal(cwd, 'gsd-code-reviewer'),

tests/init.test.cjs

@@ -1951,6 +1951,32 @@ describe('cmdInitQuick', () => {
     cleanup(tmpDir);
   });
 
+  test('init quick emits researcher_model', () => {
+    // #3936: the quick research step (Step 4.75) dispatches gsd-phase-researcher,
+    // so init quick must resolve the researcher's own model tier — parity with
+    // init plan-phase (which emits researcher_model for the same agent).
+    const result = runGsdTools('init quick "Fix login bug"', tmpDir);
+    assert.ok(result.success, `Command failed: ${result.error}`);
+
+    const output = JSON.parse(result.output);
+    assert.ok('researcher_model' in output,
+      'init quick should emit researcher_model for the Step 4.75 research dispatch');
+  });
+
+  test('init quick resolves researcher_model from model_overrides', () => {
+    fs.writeFileSync(path.join(tmpDir, '.planning', 'config.json'), JSON.stringify({
+      model_profile: 'balanced',
+      model_overrides: { 'gsd-phase-researcher': 'openai/o4-mini' },
+    }));
+
+    const result = runGsdTools('init quick "Fix login bug" --raw', tmpDir, { HOME: tmpDir });
+    assert.ok(result.success, `Command failed: ${result.error}`);
+
+    const output = JSON.parse(result.output);
+    assert.strictEqual(output.researcher_model, 'openai/o4-mini',
+      'model_overrides["gsd-phase-researcher"] must reach init quick\'s researcher_model');
+  });
+
   test('with description generates slug and task_dir with YYMMDD-xxx format', () => {
     const result = runGsdTools('init quick "Fix login bug"', tmpDir);
     assert.ok(result.success, `Command failed: ${result.error}`);

tests/quick-research.test.cjs

@@ -129,6 +129,68 @@ describe('quick workflow: research step', () => {
     );
   });
 
+  test('research step injects the researcher persona, not the planner persona', () => {
+    // #3936: the dispatch targets gsd-phase-researcher, so the persona riding the
+    // prompt and the model tier must be the researcher's own — not the planner's.
+    content = expandWorkflowSections(workflowPath);
+    const researchSection = content.substring(
+      content.indexOf('Step 4.75'),
+      content.indexOf('Step 5:')
+    );
+    assert.ok(
+      researchSection.includes('${AGENT_SKILLS_RESEARCHER}'),
+      'research step should inject the gsd-phase-researcher persona'
+    );
+    assert.ok(
+      !researchSection.includes('${AGENT_SKILLS_PLANNER}'),
+      'research step must not inject the planner persona into a researcher dispatch'
+    );
+    assert.ok(
+      researchSection.includes('model="{researcher_model}"'),
+      'research step should pin the researcher model tier, not planner_model'
+    );
+  });
+
+  test('quick workflow resolves the researcher persona', () => {
+    // #3936: AGENT_SKILLS_RESEARCHER must be resolved in quick.md's Step 2 block,
+    // the way plan-phase.md does, or the Step 4.75 interpolation expands empty.
+    const quickMd = fs.readFileSync(path.join(WORKFLOWS_DIR, 'quick.md'), 'utf8');
+    assert.ok(
+      quickMd.includes('AGENT_SKILLS_RESEARCHER=$(gsd_run query agent-skills gsd-phase-researcher)'),
+      'quick.md should resolve AGENT_SKILLS_RESEARCHER from agent-skills'
+    );
+  });
+
+  test('quick workflow parses researcher_model from init quick', () => {
+    const quickMd = fs.readFileSync(path.join(WORKFLOWS_DIR, 'quick.md'), 'utf8');
+    const parseList = quickMd.substring(
+      quickMd.indexOf('Parse JSON for:'),
+      quickMd.indexOf('```', quickMd.indexOf('Parse JSON for:'))
+    );
+    assert.ok(
+      parseList.includes('researcher_model'),
+      'quick.md parse list should include researcher_model'
+    );
+  });
+
+  test('planner and executor dispatches keep their own personas', () => {
+    // Negative space: #3936 touches only the research dispatch — the planner and
+    // executor spawns must keep their own personas and model tiers.
+    content = expandWorkflowSections(workflowPath);
+    const plannerSection = content.substring(
+      content.indexOf('Step 5: Spawn planner'),
+      content.indexOf('Step 5.5')
+    );
+    assert.ok(
+      plannerSection.includes('${AGENT_SKILLS_PLANNER}'),
+      'planner dispatch should still inject the planner persona'
+    );
+    assert.ok(
+      content.includes('${AGENT_SKILLS_EXECUTOR}'),
+      'executor dispatch should still inject the executor persona'
+    );
+  });
+
   test('research step writes RESEARCH.md', () => {
     content = expandWorkflowSections(workflowPath);
     const researchSection = content.substring(