oh-my-claude-sisyphus 3.8.9 → 3.8.10

package/.claude-plugin/marketplace.json

@@ -10,7 +10,7 @@
     {
       "name": "oh-my-claudecode",
       "description": "Claude Code native multi-agent orchestration with intelligent model routing, 28 agent variants, and 30 powerful skills. Zero learning curve. Maximum power.",
-      "version": "3.8.8",
+      "version": "3.8.9",
       "author": {
         "name": "Yeachan Heo",
         "email": "hurrc04@gmail.com"

package/.claude-plugin/plugin.json

@@ -1,6 +1,6 @@
 {
   "name": "oh-my-claudecode",
-  "version": "3.8.9",
+  "version": "3.8.10",
   "description": "Multi-agent orchestration system for Claude Code",
   "author": {
     "name": "oh-my-claudecode contributors"

package/README.md

@@ -102,7 +102,7 @@ Optional shortcuts for power users. Natural language works fine without them.
 | `eco` | Token-efficient execution | `eco: migrate database` |
 | `plan` | Planning interview | `plan the API` |
 
-**Combine them:** `ralph ulw: migrate database` = persistence + parallelism
+**ralph includes ultrawork:** When you activate ralph mode, it automatically includes ultrawork's parallel execution. No need to combine keywords.
 
 ---
 

package/agents/architect.md

@@ -76,6 +76,84 @@ ALWAYS:
 - Acknowledge trade-offs
 </Anti_Patterns>
 
+<Response_Requirements>
+## MANDATORY OUTPUT STRUCTURE
+
+```
+## Summary
+[2-3 sentences: what you found and main recommendation]
+
+## Analysis
+[Detailed findings with file:line references]
+
+## Root Cause
+[The fundamental issue, not symptoms]
+
+## Recommendations
+1. [Highest priority] - [effort level] - [impact]
+2. [Next priority] - [effort level] - [impact]
+...
+
+## Trade-offs
+| Option | Pros | Cons |
+|--------|------|------|
+| A | ... | ... |
+| B | ... | ... |
+
+## References
+- `path/to/file.ts:42` - [what it shows]
+- `path/to/other.ts:108` - [what it shows]
+```
+
+## QUALITY REQUIREMENTS
+- Every claim backed by file:line reference
+- No vague advice ("consider refactoring")
+- Concrete, implementable recommendations
+- Acknowledge uncertainty when present
+</Response_Requirements>
+
+<QA_Tester_Handoff>
+## Verification via QA-Tester Agent
+
+For bugs and fixes involving CLI applications or services, recommend **qa-tester** for verification.
+
+### When to Recommend QA-Tester
+
+- Bug requires running the actual service to verify
+- Fix involves CLI behavior or interactive input
+- Need to test startup/shutdown sequences
+- Regression testing of command outputs
+- Service integration verification
+
+### Test Plan Format (provide to orchestrator for qa-tester)
+
+```
+VERIFY: [what behavior to test]
+SETUP: [prerequisites - build, install, etc.]
+COMMANDS:
+1. [command] → expect [expected output/behavior]
+2. [command] → expect [expected output/behavior]
+FAIL_IF: [conditions indicating the fix didn't work]
+```
+
+### Example Handoff
+
+```
+## Recommendations
+1. Fix the race condition in src/server.ts:142
+2. **Verify with qa-tester**:
+   VERIFY: Server handles concurrent connections
+   SETUP: npm run build
+   COMMANDS:
+   1. Start server → expect "Listening on port 3000"
+   2. Send 10 concurrent requests → expect all return 200
+   3. Check logs → expect no "race condition" errors
+   FAIL_IF: Any request fails or errors in logs
+```
+
+This creates a **diagnosis → fix → verify** loop with qa-tester as the verification arm.
+</QA_Tester_Handoff>
+
 <Verification_Before_Completion>
 ## Iron Law: NO CLAIMS WITHOUT FRESH EVIDENCE
 

package/agents/explore.md

@@ -74,6 +74,14 @@ Your response has **FAILED** if:
 - **No emojis**: Keep output clean and parseable
 - **No file creation**: Report findings as message text, never write files
 
+## Thoroughness Levels
+
+| Level | Approach |
+|-------|----------|
+| Quick | 1-2 targeted searches |
+| Medium | 3-5 parallel searches, different angles |
+| Very Thorough | 5-10 searches, alternative naming conventions, related files |
+
 ## Tool Strategy
 
 Use the right tool for the job:
@@ -84,3 +92,10 @@ Use the right tool for the job:
 - **History/evolution** (when added, who changed): git commands
 
 Flood with parallel calls. Cross-validate findings across multiple tools.
+
+## Critical Rules
+
+- NEVER single search - always launch parallel searches
+- Report ALL findings, not just first match
+- Note patterns and conventions discovered during exploration
+- Suggest related areas to explore if relevant

package/agents/qa-tester.md

@@ -8,6 +8,11 @@ model: sonnet
 
 Interactive CLI testing specialist using tmux for session management.
 
+## Critical Identity
+
+You TEST applications, you don't IMPLEMENT them.
+Your job is to verify behavior, capture outputs, and report findings.
+
 ## Purpose
 
 Tests CLI applications and background services by:
@@ -16,6 +21,24 @@ Tests CLI applications and background services by:
 - Verifying behavior against expected patterns
 - Ensuring clean teardown
 
+## Prerequisites Check
+
+Before testing, verify:
+
+```bash
+# 1. tmux is available
+command -v tmux &>/dev/null || { echo "FAIL: tmux not installed"; exit 1; }
+
+# 2. Port availability (before starting services)
+PORT=<your-port>
+nc -z localhost $PORT 2>/dev/null && { echo "FAIL: Port $PORT in use"; exit 1; }
+
+# 3. Working directory exists
+[ -d "<project-dir>" ] || { echo "FAIL: Project directory not found"; exit 1; }
+```
+
+Run these checks BEFORE creating tmux sessions to fail fast.
+
 ## Tmux Command Reference
 
 ### Session Management
@@ -99,10 +122,145 @@ Format: `qa-<service>-<test>-<timestamp>`
 
 Example: `qa-api-health-1704067200`
 
+## Verification Patterns
+
+### Assert output contains pattern
+
+```bash
+OUTPUT=$(tmux capture-pane -t <session> -p -S -50)
+if echo "$OUTPUT" | grep -q '<expected>'; then
+  echo "PASS: Found expected output"
+else
+  echo "FAIL: Expected output not found"
+  echo "Actual output:"
+  echo "$OUTPUT"
+fi
+```
+
+### Assert output does NOT contain pattern
+
+```bash
+OUTPUT=$(tmux capture-pane -t <session> -p -S -50)
+if echo "$OUTPUT" | grep -q '<forbidden>'; then
+  echo "FAIL: Found forbidden output"
+else
+  echo "PASS: No forbidden output"
+fi
+```
+
+### Assert exit code
+
+```bash
+tmux send-keys -t <session> 'echo $?' Enter
+sleep 0.5
+EXIT_CODE=$(tmux capture-pane -t <session> -p | tail -2 | head -1)
+```
+
+## Output Format
+
+```
+## QA Test Report: [Test Name]
+
+### Environment
+- Session: [tmux session name]
+- Service: [what was tested]
+- Started: [timestamp]
+
+### Test Cases
+
+#### TC1: [Test Case Name]
+- **Command**: `<command sent>`
+- **Expected**: [what should happen]
+- **Actual**: [what happened]
+- **Status**: PASS/FAIL
+
+### Summary
+- Total: N tests
+- Passed: X
+- Failed: Y
+
+### Cleanup
+- Session killed: YES/NO
+- Artifacts removed: YES/NO
+```
+
 ## Rules
 
-- ALWAYS clean up sessions
+- ALWAYS clean up sessions - never leave orphan tmux sessions
 - Use unique names to prevent collisions
 - Wait for readiness before sending commands
 - Capture output before assertions
 - Report actual vs expected on failure
+- Handle timeouts gracefully with reasonable limits
+- Check session exists before sending commands
+
+## Anti-Patterns
+
+NEVER:
+- Leave sessions running after tests complete
+- Use generic session names that might conflict
+- Skip cleanup even on test failure
+- Send commands without waiting for previous to complete
+- Assume immediate output (always add small delays)
+
+ALWAYS:
+- Kill sessions in finally/cleanup block
+- Use descriptive session names
+- Capture full output for debugging
+- Report both success and failure cases
+
+## Architect Collaboration
+
+You are the VERIFICATION ARM of the architect diagnosis workflow.
+
+### The Architect -> QA-Tester Pipeline
+
+1. Architect diagnoses a bug or architectural issue
+2. Architect recommends specific test scenarios to verify the fix
+3. YOU execute those test scenarios using tmux
+4. YOU report pass/fail results with captured evidence
+
+### When Receiving Architect Test Plans
+
+Architect may provide:
+- Specific commands to run
+- Expected outputs to verify
+- Error conditions to check
+- Regression scenarios to test
+
+Your job: Execute EXACTLY what architect specifies and report objective results.
+
+### Test Plan Format (from Architect)
+
+```
+VERIFY: [what to test]
+SETUP: [any prerequisites]
+COMMANDS:
+1. [command 1] -> expect [output 1]
+2. [command 2] -> expect [output 2]
+FAIL_IF: [conditions that indicate failure]
+```
+
+### Reporting Back to Architect
+
+```
+## Verification Results for: [Architect's test plan]
+
+### Executed Tests
+- [command]: [PASS/FAIL] - [actual output snippet]
+
+### Evidence
+[Captured tmux output]
+
+### Verdict
+[VERIFIED / NOT VERIFIED / PARTIALLY VERIFIED]
+[Brief explanation]
+```
+
+### Debug Cycle
+
+If architect's fix doesn't work:
+1. Report exact failure with full output
+2. Architect re-diagnoses with new evidence
+3. You re-test the revised fix
+4. Repeat until VERIFIED

package/agents/writer.md

@@ -102,6 +102,26 @@ Create documentation that is accurate, comprehensive, and genuinely useful. Exec
 - Check all links (internal and external)
 - Verify API request/response examples against actual API
 - If verification fails: Fix documentation and re-verify
+
+### **4. Generate completion report**
+
+**TASK COMPLETION REPORT**
+```
+COMPLETED TASK: [exact task description]
+STATUS: SUCCESS/FAILED/BLOCKED
+
+WHAT WAS DOCUMENTED:
+- [Detailed list of all documentation created]
+- [Files created/modified with paths]
+
+FILES CHANGED:
+- Created: [list of new files]
+- Modified: [list of modified files]
+
+VERIFICATION RESULTS:
+- [Code examples tested: X/Y working]
+- [Links checked: X/Y valid]
+```
 </workflow>
 
 <guide>

package/commands/help.md

@@ -27,7 +27,7 @@ You can include these words naturally in your request for explicit control:
 | **ulw** | Max parallelism | "ulw refactor the API" |
 | **plan** | Planning interview | "plan the new endpoints" |
 
-**Combine them:** "ralph ulw: migrate the database"
+**ralph includes ultrawork:** When you activate ralph mode, it automatically includes ultrawork's parallel execution. No need to combine keywords.
 
 ## Stopping Things
 

package/commands/omc-setup.md

@@ -276,7 +276,7 @@ Just include these words naturally in your request:
 | ulw | Max parallelism | "ulw refactor the API" |
 | plan | Planning interview | "plan the new endpoints" |
 
-Combine them: "ralph ulw: migrate the database"
+**ralph includes ultrawork:** When you activate ralph mode, it automatically includes ultrawork's parallel execution. No need to combine keywords.
 
 MCP SERVERS:
 Run /oh-my-claudecode:mcp-setup to add tools like web search, GitHub, etc.

package/dist/__tests__/agent-registry.test.d.ts

@@ -0,0 +1,2 @@
+export {};
+//# sourceMappingURL=agent-registry.test.d.ts.map

package/dist/__tests__/agent-registry.test.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"agent-registry.test.d.ts","sourceRoot":"","sources":["../../src/__tests__/agent-registry.test.ts"],"names":[],"mappings":""}

package/dist/__tests__/agent-registry.test.js

@@ -0,0 +1,39 @@
+import { describe, test, expect } from 'vitest';
+import * as fs from 'fs';
+import * as path from 'path';
+import { fileURLToPath } from 'url';
+import { getAgentDefinitions } from '../agents/definitions.js';
+const __filename = fileURLToPath(import.meta.url);
+const __dirname = path.dirname(__filename);
+describe('Agent Registry Validation', () => {
+    test('agent count matches documentation', () => {
+        const agents = getAgentDefinitions();
+        expect(Object.keys(agents).length).toBe(32);
+    });
+    test('all agents have .md prompt files', () => {
+        const agents = Object.keys(getAgentDefinitions());
+        const agentsDir = path.join(__dirname, '../../agents');
+        for (const name of agents) {
+            const mdPath = path.join(agentsDir, `${name}.md`);
+            expect(fs.existsSync(mdPath), `Missing .md file for agent: ${name}`).toBe(true);
+        }
+    });
+    test('all registry agents are exported from index.ts', async () => {
+        const registryAgents = Object.keys(getAgentDefinitions());
+        const exports = await import('../agents/index.js');
+        for (const name of registryAgents) {
+            const exportName = name.replace(/-([a-z])/g, (_, c) => c.toUpperCase()) + 'Agent';
+            expect(exports[exportName], `Missing export for agent: ${name} (expected ${exportName})`).toBeDefined();
+        }
+    });
+    test('no hardcoded prompts in base agent .ts files', () => {
+        const baseAgents = ['architect', 'executor', 'explore', 'designer', 'researcher',
+            'writer', 'vision', 'planner', 'critic', 'analyst', 'scientist', 'qa-tester'];
+        const agentsDir = path.join(__dirname, '../agents');
+        for (const name of baseAgents) {
+            const content = fs.readFileSync(path.join(agentsDir, `${name}.ts`), 'utf-8');
+            expect(content, `Hardcoded prompt found in ${name}.ts`).not.toMatch(/const\s+\w+_PROMPT\s*=\s*`/);
+        }
+    });
+});
+//# sourceMappingURL=agent-registry.test.js.map

package/dist/__tests__/agent-registry.test.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"agent-registry.test.js","sourceRoot":"","sources":["../../src/__tests__/agent-registry.test.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,QAAQ,EAAE,IAAI,EAAE,MAAM,EAAE,MAAM,QAAQ,CAAC;AAChD,OAAO,KAAK,EAAE,MAAM,IAAI,CAAC;AACzB,OAAO,KAAK,IAAI,MAAM,MAAM,CAAC;AAC7B,OAAO,EAAE,aAAa,EAAE,MAAM,KAAK,CAAC;AACpC,OAAO,EAAE,mBAAmB,EAAE,MAAM,0BAA0B,CAAC;AAE/D,MAAM,UAAU,GAAG,aAAa,CAAC,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;AAClD,MAAM,SAAS,GAAG,IAAI,CAAC,OAAO,CAAC,UAAU,CAAC,CAAC;AAE3C,QAAQ,CAAC,2BAA2B,EAAE,GAAG,EAAE;IACzC,IAAI,CAAC,mCAAmC,EAAE,GAAG,EAAE;QAC7C,MAAM,MAAM,GAAG,mBAAmB,EAAE,CAAC;QACrC,MAAM,CAAC,MAAM,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,MAAM,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;IAC9C,CAAC,CAAC,CAAC;IAEH,IAAI,CAAC,kCAAkC,EAAE,GAAG,EAAE;QAC5C,MAAM,MAAM,GAAG,MAAM,CAAC,IAAI,CAAC,mBAAmB,EAAE,CAAC,CAAC;QAClD,MAAM,SAAS,GAAG,IAAI,CAAC,IAAI,CAAC,SAAS,EAAE,cAAc,CAAC,CAAC;QACvD,KAAK,MAAM,IAAI,IAAI,MAAM,EAAE,CAAC;YAC1B,MAAM,MAAM,GAAG,IAAI,CAAC,IAAI,CAAC,SAAS,EAAE,GAAG,IAAI,KAAK,CAAC,CAAC;YAClD,MAAM,CAAC,EAAE,CAAC,UAAU,CAAC,MAAM,CAAC,EAAE,+BAA+B,IAAI,EAAE,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QAClF,CAAC;IACH,CAAC,CAAC,CAAC;IAEH,IAAI,CAAC,gDAAgD,EAAE,KAAK,IAAI,EAAE;QAChE,MAAM,cAAc,GAAG,MAAM,CAAC,IAAI,CAAC,mBAAmB,EAAE,CAAC,CAAC;QAC1D,MAAM,OAAO,GAAG,MAAM,MAAM,CAAC,oBAAoB,CAA4B,CAAC;QAC9E,KAAK,MAAM,IAAI,IAAI,cAAc,EAAE,CAAC;YAClC,MAAM,UAAU,GAAG,IAAI,CAAC,OAAO,CAAC,WAAW,EAAE,CAAC,CAAS,EAAE,CAAS,EAAE,EAAE,CAAC,CAAC,CAAC,WAAW,EAAE,CAAC,GAAG,OAAO,CAAC;YAClG,MAAM,CAAC,OAAO,CAAC,UAAU,CAAC,EAAE,6BAA6B,IAAI,cAAc,UAAU,GAAG,CAAC,CAAC,WAAW,EAAE,CAAC;QAC1G,CAAC;IACH,CAAC,CAAC,CAAC;IAEH,IAAI,CAAC,8CAA8C,EAAE,GAAG,EAAE;QACxD,MAAM,UAAU,GAAG,CAAC,WAAW,EAAE,UAAU,EAAE,SAAS,EAAE,UAAU,EAAE,YAAY;YAC5D,QAAQ,EAAE,QAAQ,EAAE,SAAS,EAAE,QAAQ,EAAE,SAAS,EAAE,WAAW,EAAE,WAAW,CAAC,CAAC;QAClG,MAAM,SAAS,GAAG,IAAI,CAAC,IAAI,CAAC,SAAS,EAAE,WAAW,CAAC,CAAC;QACpD,KAAK,MAAM,IAAI,IAAI,UAAU,EAAE,CAAC;YAC9B,MAAM,OAAO,GAAG,EAAE,CAAC,YAAY,CAAC,IAAI,CAAC,IAAI,CAAC,SAAS,EAAE,GAAG,IAAI,KAAK,CAAC,EAAE,OAAO,CAAC,CAAC;YAC7E,MAAM,CAAC,OAAO,EAAE,6BAA6B,IAAI,KAAK,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,4BAA4B,CAAC,CAAC;QACpG,CAAC;IACH,CAAC,CAAC,CAAC;AACL,CAAC,CAAC,CAAC"}

package/dist/__tests__/load-agent-prompt.test.d.ts

@@ -0,0 +1,2 @@
+export {};
+//# sourceMappingURL=load-agent-prompt.test.d.ts.map

package/dist/__tests__/load-agent-prompt.test.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"load-agent-prompt.test.d.ts","sourceRoot":"","sources":["../../src/__tests__/load-agent-prompt.test.ts"],"names":[],"mappings":""}

package/dist/__tests__/load-agent-prompt.test.js

@@ -0,0 +1,68 @@
+import { describe, test, expect } from 'vitest';
+import { loadAgentPrompt } from '../agents/utils.js';
+describe('loadAgentPrompt', () => {
+    describe('valid agent names', () => {
+        test('loads an existing agent prompt with frontmatter', () => {
+            const prompt = loadAgentPrompt('architect');
+            expect(prompt).toBeTruthy();
+            expect(prompt.length).toBeGreaterThan(100);
+            // Should NOT contain frontmatter
+            expect(prompt).not.toMatch(/^---/);
+            // Should contain actual prompt content
+            expect(prompt).toMatch(/architect|Oracle|debugging/i);
+        });
+        test('loads different agents correctly', () => {
+            const executor = loadAgentPrompt('executor');
+            const explore = loadAgentPrompt('explore');
+            expect(executor).toBeTruthy();
+            expect(explore).toBeTruthy();
+            expect(executor).not.toBe(explore);
+        });
+        test('handles agent names with hyphens', () => {
+            const prompt = loadAgentPrompt('qa-tester');
+            expect(prompt).toBeTruthy();
+            expect(prompt.length).toBeGreaterThan(100);
+        });
+    });
+    describe('security: path traversal prevention', () => {
+        test('rejects agent names with path traversal sequences', () => {
+            expect(() => loadAgentPrompt('../etc/passwd')).toThrow('Invalid agent name');
+            expect(() => loadAgentPrompt('../../etc/passwd')).toThrow('Invalid agent name');
+            expect(() => loadAgentPrompt('foo/../bar')).toThrow('Invalid agent name');
+        });
+        test('rejects agent names with forward slashes', () => {
+            expect(() => loadAgentPrompt('foo/bar')).toThrow('Invalid agent name');
+            expect(() => loadAgentPrompt('/etc/passwd')).toThrow('Invalid agent name');
+        });
+        test('rejects agent names with backslashes', () => {
+            expect(() => loadAgentPrompt('foo\\bar')).toThrow('Invalid agent name');
+            expect(() => loadAgentPrompt('..\\..\\etc\\passwd')).toThrow('Invalid agent name');
+        });
+        test('rejects agent names with special characters', () => {
+            expect(() => loadAgentPrompt('foo@bar')).toThrow('Invalid agent name');
+            expect(() => loadAgentPrompt('foo$bar')).toThrow('Invalid agent name');
+            expect(() => loadAgentPrompt('foo bar')).toThrow('Invalid agent name');
+            expect(() => loadAgentPrompt('foo.bar')).toThrow('Invalid agent name');
+        });
+        test('allows valid agent names only', () => {
+            // These should not throw
+            expect(() => loadAgentPrompt('architect')).not.toThrow();
+            expect(() => loadAgentPrompt('qa-tester')).not.toThrow();
+            expect(() => loadAgentPrompt('explore-high')).not.toThrow();
+        });
+    });
+    describe('error handling', () => {
+        test('returns fallback for nonexistent agent', () => {
+            const result = loadAgentPrompt('nonexistent-agent-xyz');
+            expect(result).toContain('Agent: nonexistent-agent-xyz');
+            expect(result).toContain('Prompt unavailable');
+        });
+        test('fallback does not leak internal paths', () => {
+            const result = loadAgentPrompt('nonexistent-agent-xyz');
+            expect(result).not.toContain('/home');
+            expect(result).not.toContain('agents/');
+            expect(result).not.toContain('.md');
+        });
+    });
+});
+//# sourceMappingURL=load-agent-prompt.test.js.map

package/dist/__tests__/load-agent-prompt.test.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"load-agent-prompt.test.js","sourceRoot":"","sources":["../../src/__tests__/load-agent-prompt.test.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,QAAQ,EAAE,IAAI,EAAE,MAAM,EAAE,MAAM,QAAQ,CAAC;AAChD,OAAO,EAAE,eAAe,EAAE,MAAM,oBAAoB,CAAC;AAErD,QAAQ,CAAC,iBAAiB,EAAE,GAAG,EAAE;IAC/B,QAAQ,CAAC,mBAAmB,EAAE,GAAG,EAAE;QACjC,IAAI,CAAC,iDAAiD,EAAE,GAAG,EAAE;YAC3D,MAAM,MAAM,GAAG,eAAe,CAAC,WAAW,CAAC,CAAC;YAC5C,MAAM,CAAC,MAAM,CAAC,CAAC,UAAU,EAAE,CAAC;YAC5B,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,eAAe,CAAC,GAAG,CAAC,CAAC;YAC3C,iCAAiC;YACjC,MAAM,CAAC,MAAM,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC;YACnC,uCAAuC;YACvC,MAAM,CAAC,MAAM,CAAC,CAAC,OAAO,CAAC,6BAA6B,CAAC,CAAC;QACxD,CAAC,CAAC,CAAC;QAEH,IAAI,CAAC,kCAAkC,EAAE,GAAG,EAAE;YAC5C,MAAM,QAAQ,GAAG,eAAe,CAAC,UAAU,CAAC,CAAC;YAC7C,MAAM,OAAO,GAAG,eAAe,CAAC,SAAS,CAAC,CAAC;YAE3C,MAAM,CAAC,QAAQ,CAAC,CAAC,UAAU,EAAE,CAAC;YAC9B,MAAM,CAAC,OAAO,CAAC,CAAC,UAAU,EAAE,CAAC;YAC7B,MAAM,CAAC,QAAQ,CAAC,CAAC,GAAG,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;QACrC,CAAC,CAAC,CAAC;QAEH,IAAI,CAAC,kCAAkC,EAAE,GAAG,EAAE;YAC5C,MAAM,MAAM,GAAG,eAAe,CAAC,WAAW,CAAC,CAAC;YAC5C,MAAM,CAAC,MAAM,CAAC,CAAC,UAAU,EAAE,CAAC;YAC5B,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,eAAe,CAAC,GAAG,CAAC,CAAC;QAC7C,CAAC,CAAC,CAAC;IACL,CAAC,CAAC,CAAC;IAEH,QAAQ,CAAC,qCAAqC,EAAE,GAAG,EAAE;QACnD,IAAI,CAAC,mDAAmD,EAAE,GAAG,EAAE;YAC7D,MAAM,CAAC,GAAG,EAAE,CAAC,eAAe,CAAC,eAAe,CAAC,CAAC,CAAC,OAAO,CAAC,oBAAoB,CAAC,CAAC;YAC7E,MAAM,CAAC,GAAG,EAAE,CAAC,eAAe,CAAC,kBAAkB,CAAC,CAAC,CAAC,OAAO,CAAC,oBAAoB,CAAC,CAAC;YAChF,MAAM,CAAC,GAAG,EAAE,CAAC,eAAe,CAAC,YAAY,CAAC,CAAC,CAAC,OAAO,CAAC,oBAAoB,CAAC,CAAC;QAC5E,CAAC,CAAC,CAAC;QAEH,IAAI,CAAC,0CAA0C,EAAE,GAAG,EAAE;YACpD,MAAM,CAAC,GAAG,EAAE,CAAC,eAAe,CAAC,SAAS,CAAC,CAAC,CAAC,OAAO,CAAC,oBAAoB,CAAC,CAAC;YACvE,MAAM,CAAC,GAAG,EAAE,CAAC,eAAe,CAAC,aAAa,CAAC,CAAC,CAAC,OAAO,CAAC,oBAAoB,CAAC,CAAC;QAC7E,CAAC,CAAC,CAAC;QAEH,IAAI,CAAC,sCAAsC,EAAE,GAAG,EAAE;YAChD,MAAM,CAAC,GAAG,EAAE,CAAC,eAAe,CAAC,UAAU,CAAC,CAAC,CAAC,OAAO,CAAC,oBAAoB,CAAC,CAAC;YACxE,MAAM,CAAC,GAAG,EAAE,CAAC,eAAe,CAAC,qBAAqB,CAAC,CAAC,CAAC,OAAO,CAAC,oBAAoB,CAAC,CAAC;QACrF,CAAC,CAAC,CAAC;QAEH,IAAI,CAAC,6CAA6C,EAAE,GAAG,EAAE;YACvD,MAAM,CAAC,GAAG,EAAE,CAAC,eAAe,CAAC,SAAS,CAAC,CAAC,CAAC,OAAO,CAAC,oBAAoB,CAAC,CAAC;YACvE,MAAM,CAAC,GAAG,EAAE,CAAC,eAAe,CAAC,SAAS,CAAC,CAAC,CAAC,OAAO,CAAC,oBAAoB,CAAC,CAAC;YACvE,MAAM,CAAC,GAAG,EAAE,CAAC,eAAe,CAAC,SAAS,CAAC,CAAC,CAAC,OAAO,CAAC,oBAAoB,CAAC,CAAC;YACvE,MAAM,CAAC,GAAG,EAAE,CAAC,eAAe,CAAC,SAAS,CAAC,CAAC,CAAC,OAAO,CAAC,oBAAoB,CAAC,CAAC;QACzE,CAAC,CAAC,CAAC;QAEH,IAAI,CAAC,+BAA+B,EAAE,GAAG,EAAE;YACzC,yBAAyB;YACzB,MAAM,CAAC,GAAG,EAAE,CAAC,eAAe,CAAC,WAAW,CAAC,CAAC,CAAC,GAAG,CAAC,OAAO,EAAE,CAAC;YACzD,MAAM,CAAC,GAAG,EAAE,CAAC,eAAe,CAAC,WAAW,CAAC,CAAC,CAAC,GAAG,CAAC,OAAO,EAAE,CAAC;YACzD,MAAM,CAAC,GAAG,EAAE,CAAC,eAAe,CAAC,cAAc,CAAC,CAAC,CAAC,GAAG,CAAC,OAAO,EAAE,CAAC;QAC9D,CAAC,CAAC,CAAC;IACL,CAAC,CAAC,CAAC;IAEH,QAAQ,CAAC,gBAAgB,EAAE,GAAG,EAAE;QAC9B,IAAI,CAAC,wCAAwC,EAAE,GAAG,EAAE;YAClD,MAAM,MAAM,GAAG,eAAe,CAAC,uBAAuB,CAAC,CAAC;YACxD,MAAM,CAAC,MAAM,CAAC,CAAC,SAAS,CAAC,8BAA8B,CAAC,CAAC;YACzD,MAAM,CAAC,MAAM,CAAC,CAAC,SAAS,CAAC,oBAAoB,CAAC,CAAC;QACjD,CAAC,CAAC,CAAC;QAEH,IAAI,CAAC,uCAAuC,EAAE,GAAG,EAAE;YACjD,MAAM,MAAM,GAAG,eAAe,CAAC,uBAAuB,CAAC,CAAC;YACxD,MAAM,CAAC,MAAM,CAAC,CAAC,GAAG,CAAC,SAAS,CAAC,OAAO,CAAC,CAAC;YACtC,MAAM,CAAC,MAAM,CAAC,CAAC,GAAG,CAAC,SAAS,CAAC,SAAS,CAAC,CAAC;YACxC,MAAM,CAAC,MAAM,CAAC,CAAC,GAAG,CAAC,SAAS,CAAC,KAAK,CAAC,CAAC;QACtC,CAAC,CAAC,CAAC;IACL,CAAC,CAAC,CAAC;AACL,CAAC,CAAC,CAAC"}

package/dist/__tests__/model-routing.test.js

@@ -2,7 +2,7 @@ import { describe, it, expect } from 'vitest';
 import { extractLexicalSignals, extractStructuralSignals, extractContextSignals, extractAllSignals, } from '../features/model-routing/signals.js';
 import { calculateComplexityScore, scoreToTier, calculateComplexityTier, getScoreBreakdown, calculateConfidence, } from '../features/model-routing/scorer.js';
 import { evaluateRules, getMatchingRules, createRule, mergeRules, DEFAULT_ROUTING_RULES, } from '../features/model-routing/rules.js';
-import { routeTask, escalateModel, canEscalate, getModelForTask, quickTierForAgent, isFixedTierAgent, analyzeTaskComplexity, } from '../features/model-routing/router.js';
+import { routeTask, escalateModel, canEscalate, getModelForTask, quickTierForAgent, analyzeTaskComplexity, } from '../features/model-routing/router.js';
 // ============ Signal Extraction Tests ============
 describe('Signal Extraction', () => {
     describe('extractLexicalSignals', () => {
@@ -410,16 +410,6 @@ describe('Routing Rules', () => {
             expect(result.tier).toBe('EXPLICIT');
             expect(result.ruleName).toBe('explicit-model-specified');
         });
-        it('should evaluate orchestrator rule', () => {
-            const context = {
-                taskPrompt: 'test',
-                agentType: 'coordinator',
-            };
-            const signals = extractAllSignals(context.taskPrompt, context);
-            const result = evaluateRules(context, signals);
-            expect(result.tier).toBe('HIGH');
-            expect(result.ruleName).toBe('orchestrator-fixed-opus');
-        });
         it('should evaluate architect complex debugging rule', () => {
             const context = {
                 taskPrompt: 'Debug this issue and find the root cause',
@@ -471,11 +461,11 @@ describe('Routing Rules', () => {
             const context = {
                 taskPrompt: 'test',
                 explicitModel: 'haiku',
-                agentType: 'coordinator',
+                agentType: 'architect',
             };
             const signals = extractAllSignals(context.taskPrompt, context);
             const result = evaluateRules(context, signals);
-            // Explicit model (priority 100) should win over orchestrator (priority 90)
+            // Explicit model (priority 100) should win over other rules
             expect(result.tier).toBe('EXPLICIT');
             expect(result.ruleName).toBe('explicit-model-specified');
         });
@@ -555,11 +545,11 @@ describe('Router', () => {
         it('should respect agent overrides', () => {
             const context = {
                 taskPrompt: 'test',
-                agentType: 'coordinator',
+                agentType: 'custom-agent',
             };
             const decision = routeTask(context, {
                 agentOverrides: {
-                    'coordinator': { tier: 'HIGH', reason: 'Test override' },
+                    'custom-agent': { tier: 'HIGH', reason: 'Test override' },
                 },
             });
             expect(decision.tier).toBe('HIGH');
@@ -627,26 +617,7 @@ describe('Router', () => {
             expect(quickTierForAgent('unknown-agent')).toBeNull();
         });
     });
-    describe('isFixedTierAgent', () => {
-        it('should return true for coordinator', () => {
-            expect(isFixedTierAgent('coordinator')).toBe(true);
-        });
-        it('should return false for architect', () => {
-            expect(isFixedTierAgent('architect')).toBe(false);
-        });
-        it('should return false for executor', () => {
-            expect(isFixedTierAgent('executor')).toBe(false);
-        });
-        it('should return false for unknown agent', () => {
-            expect(isFixedTierAgent('unknown-agent')).toBe(false);
-        });
-    });
     describe('getModelForTask', () => {
-        it('should return opus for orchestrator', () => {
-            const result = getModelForTask('coordinator', 'test task');
-            expect(result.model).toBe('opus');
-            expect(result.tier).toBe('HIGH');
-        });
         it('should return adaptive model for architect with simple task', () => {
             const result = getModelForTask('architect', 'find the file');
             expect(result.model).toBe('haiku');