Menu
Multi-agent parallel investigation for complex VCV Rack problems
Instalar com Codex ou Claude Copie este prompt, cole no Codex, Claude ou outro assistente e deixe que ele revise a página da skill e instale para você.
Baseado na classificação ocupacional SOC
Build coordination wrapper for VCV Rack modules using Makefile
Load module context from handoff files to resume work
Validate panel ↔ creative brief consistency, catch design drift
Adaptive brainstorming for VCV Rack module concepts and improvements
Version management, bug fixes, feature additions for VCV Rack modules
Manage complete module lifecycle - install, uninstall, reset, destroy
| name | deep-research |
| description | Multi-agent parallel investigation for complex VCV Rack problems |
| model | claude-opus-4-1-20250805 |
| allowed-tools | ["Read","Grep","Task","WebSearch"] |
| preconditions | ["Problem is non-trivial (quick lookup insufficient)"] |
| extended-thinking | false |
| timeout | 3600 |
Purpose: Multi-level autonomous investigation for complex VCV Rack module development problems using graduated research depth protocol.
This skill performs thorough research on technical problems by consulting multiple knowledge sources with automatic escalation based on confidence. Starts fast (5 min local lookup) and escalates as needed (up to 60 min parallel investigation).
Why graduated protocol matters:
Most problems have known solutions (Level 1: 5 min). Complex problems need deep research (Level 3: 60 min). Graduated protocol prevents over-researching simple problems while ensuring complex problems get thorough investigation.
Key innovation: Stops as soon as confident answer found. User controls depth via decision menus.
CRITICAL: This skill is read-only and advisory only. It NEVER:
Workflow handoff:
Note: The routing instruction is a directive to the main conversation. When the orchestrator sees "Invoke module-improve skill", it will use the Skill tool to invoke module-improve. This is the standard handoff protocol.
Why separation matters:
Invoked by:
/research [topic]Entry parameters:
Goal: Find solution in local knowledge base or basic Rack SDK docs
Parse research topic/question
Search local troubleshooting docs:
# Search by keywords
grep -r "[topic keywords]" troubleshooting/
# Check relevant categories
ls troubleshooting/[category]/
# Search for module-specific or general issues
Check Rack SDK documentation for direct answers:
Assess confidence:
Decision point:
If HIGH confidence:
✓ Level 1 complete (found solution)
Solution: [Brief description]
Source: [Local doc or Rack SDK API]
What's next?
1. Apply solution (recommended)
2. Review details - See full explanation
3. Continue deeper - Escalate to Level 2 for more options
4. Other
If MEDIUM/LOW confidence: Auto-escalate to Level 2 with notification:
Level 1: No confident solution found
Escalating to Level 2 (moderate investigation)...
Goal: Find authoritative answer through comprehensive documentation and community knowledge
Rack SDK deep-dive:
VCV Rack community forum search via WebSearch:
Query: "site:community.vcvrack.com [topic]"
Look for:
- Recent threads (last 2 years)
- Accepted solutions
- VCV team responses
- Multiple user confirmations
GitHub issue search:
Query: "site:github.com/VCVRack/Rack [topic]"
Look for:
- Closed issues with solutions
- Pull requests addressing problem
- Code examples in discussions
- Version-specific fixes
Synthesize findings:
Decision point:
If HIGH/MEDIUM confidence:
✓ Level 2 complete (found N solutions)
Investigated sources:
- Rack SDK docs: [Finding]
- VCV Community: [N threads]
- GitHub: [N issues]
Solutions found:
1. [Solution 1] (recommended)
2. [Solution 2] (alternative)
What's next?
1. Apply recommended solution
2. Review all options - Compare approaches
3. Continue deeper - Escalate to Level 3 (parallel investigation)
4. Other
If LOW confidence OR novel problem: Auto-escalate to Level 3 with notification:
Level 2: Complex problem detected (requires original implementation)
Escalating to Level 3 (deep research with parallel investigation)...
Switching to Opus model + extended thinking (may take up to 60 min)
Goal: Comprehensive investigation with parallel subagent research for novel/complex problems
Switch to Opus model with extended thinking:
extended_thinking: true
thinking_budget: 15000
timeout: 3600 # 60 min
Identify research approaches:
Analyze problem and determine 2-3 investigation paths:
For DSP algorithm questions:
For novel feature implementation:
For performance optimization:
Spawn parallel research subagents:
Use Task tool to spawn 2-3 specialized research agents in fresh contexts:
// Example: CV scaling research for V/Oct tracking
Task(
(subagent_type = "general-purpose"),
(description = "Research V/Oct standard"),
(prompt = `
Investigate 1V/octave CV standard for pitch tracking in VCV Rack.
Research:
1. Voltage-to-frequency conversion formula
2. VCV Rack API support (dsp::FREQ_* constants?)
3. Implementation complexity (1-5 scale)
4. Common pitfalls and edge cases
5. Reference implementations
6. Code examples
Return structured findings:
{
"approach": "V/Oct conversion",
"rack_support": "description",
"complexity": 2,
"formula": "...",
"pros": ["...", "..."],
"cons": ["...", "..."],
"references": ["...", "..."]
}
`)
);
Task(
(subagent_type = "general-purpose"),
(description = "Research exponential FM"),
(prompt = `
Investigate exponential frequency modulation for VCV Rack oscillators.
Research:
1. Linear vs exponential FM differences
2. VCV Rack conventions (how much FM range?)
3. Implementation complexity (1-5 scale)
4. CPU performance characteristics
5. Code examples or references
Return structured findings:
{
"approach": "Exponential FM",
"rack_conventions": "description",
"complexity": 3,
"typical_range": "...",
"pros": ["...", "..."],
"cons": ["...", "..."],
"references": ["...", "..."]
}
`)
);
Task(
(subagent_type = "general-purpose"),
(description = "Research polyphony implementation"),
(prompt = `
Research polyphonic voice allocation for VCV Rack modules.
Investigate:
1. VCV Rack polyphony API (16 channels max)
2. Voice allocation strategies
3. Common patterns from existing modules
4. Performance considerations
Return structured findings:
{
"approach": "Polyphony",
"max_channels": 16,
"allocation_strategy": "...",
"rack_patterns": "...",
"references": ["...", "..."]
}
`)
);
Each subagent:
Main context synthesis (extended thinking):
After all subagents return, use extended thinking to synthesize:
Academic paper search (if applicable):
For DSP algorithms, search for authoritative papers:
Generate comprehensive report:
Use structured format with populated values:
{
"agent": "deep-research",
"status": "success",
"level": 3,
"topic": "[research topic]",
"findings": [
{
"solution": "[Approach 1]",
"confidence": "high|medium|low",
"description": "...",
"pros": ["...", "..."],
"cons": ["...", "..."],
"implementation_complexity": 3,
"references": ["...", "..."]
},
{
"solution": "[Approach 2]",
...
}
],
"recommendation": 0, // Index into findings
"reasoning": "Why this approach is recommended...",
"sources": [
"Rack SDK: ...",
"VCV Community: ...",
"Paper: ...",
"Subagent findings: ..."
],
"time_spent_minutes": 45,
"escalation_path": [1, 2, 3]
}
Present findings:
✓ Level 3 complete (comprehensive investigation)
Investigated N approaches:
1. [Approach 1] - Complexity: 3/5, CPU: Low, Quality: High
2. [Approach 2] - Complexity: 2/5, CPU: Medium, Quality: High
3. [Approach 3] - Complexity: 4/5, CPU: Low, Quality: Very High
Recommendation: [Approach 2]
Reasoning: [Why this is the best fit]
What's next?
1. Apply recommended solution (recommended)
2. Review all findings - See detailed comparison
3. Try alternative approach - [Approach N name]
4. Document findings - Save to troubleshooting/
5. Other
All levels return structured report. Format depends on level:
## Research Report: [Topic]
**Level:** 1 (Quick Check)
**Time:** 5 minutes
**Confidence:** HIGH
**Source:** troubleshooting/[category]/[file].md
**Solution:**
[Direct answer from local docs or Rack SDK API]
**How to apply:**
[Step-by-step]
**Reference:** [Source file path or Rack SDK API link]
## Research Report: [Topic]
**Level:** 2 (Moderate Investigation)
**Time:** 18 minutes
**Confidence:** MEDIUM-HIGH
**Sources:**
- Rack SDK docs: [Finding]
- VCV Community: 3 threads reviewed
- GitHub: 2 issues reviewed
**Solutions:**
### Option 1: [Recommended]
**Approach:** ...
**Pros:** ...
**Cons:** ...
**Implementation:** ...
### Option 2: [Alternative]
**Approach:** ...
**Pros:** ...
**Cons:** ...
**Implementation:** ...
**Recommendation:** Option 1
**Reasoning:** [Why recommended]
**References:**
- [Source 1]
- [Source 2]
## Research Report: [Topic]
**Level:** 3 (Deep Research)
**Time:** 45 minutes
**Confidence:** HIGH
**Model:** Opus + Extended Thinking
**Investigation approach:**
- Spawned 3 parallel research subagents
- Investigated 3 distinct approaches
- Synthesized findings with extended thinking
**Findings:**
### Approach 1: [Name]
**Description:** ...
**VCV Rack support:** [API/pattern details]
**Complexity:** 2/5
**CPU cost:** Medium
**Quality:** High
**Pros:**
- Works for all cases
- VCV Rack native API
**Cons:**
- Higher CPU than alternatives
**References:**
- Rack SDK documentation
- Subagent findings
### Approach 2: [Name]
**Description:** ...
[Same structure]
### Approach 3: [Name]
**Description:** ...
[Same structure]
**Recommendation:** Approach 1
**Reasoning:**
[Detailed explanation of why this approach is best for the use case]
**Implementation Roadmap:**
1. [Step 1 with Rack API details]
2. [Step 2 with code pattern]
3. [Step 3 with integration notes]
4. [Step 4 with testing approach]
**Sources:**
- Rack SDK: [specific API documentation]
- Paper/Reference: [if applicable]
- VCV Community: [relevant threads]
- Subagent research: 3 parallel investigations
**Testing strategy:**
- [Test 1 description]
- [Test 2 description]
- [Validation approach]
After each level, present decision menu (user controls depth):
What's next?
1. Apply solution (recommended) - [One-line description]
2. Review details - See full explanation
3. Continue deeper - Escalate to Level 2 for more options
4. Other
Handle responses:
IMPORTANT: This skill is read-only. Never edit code, run builds, or modify files. All implementation happens through module-improve skill after research completes.
What's next?
1. Apply recommended solution - [Solution name]
2. Review all options - Compare N approaches
3. Continue deeper - Escalate to Level 3 (may take up to 60 min)
4. Other
Handle responses:
What's next?
1. Apply recommended solution (recommended) - [Solution name]
2. Review all findings - See detailed comparison
3. Try alternative approach - [Alternative name]
4. Document findings - Save to troubleshooting/ knowledge base
5. Other
Handle responses:
troubleshooter Level 4:
When troubleshooter agent exhausts Levels 0-3, it invokes deep-research:
## troubleshooter - Level 4
If Levels 0-3 insufficient, escalate to deep-research skill:
"I need to investigate this more deeply. Invoking deep-research skill..."
[Invoke deep-research with problem context]
deep-research handles:
- Graduated research protocol (Levels 1-3)
- Parallel investigation (Level 3)
- Extended thinking budget
- Returns structured report with recommendations
Integration flow:
After successful application of Level 2-3 findings:
Auto-suggest documentation:
✓ Solution applied successfully
This was a complex problem (Level N research). Document for future reference?
1. Yes - Create troubleshooting doc (recommended)
2. No - Skip documentation
3. Other
If user chooses "Yes":
The feedback loop:
Research is successful when:
Level-specific:
Level 1:
Level 2:
Level 3:
Timeout (>60 min):
⚠️ Research timeout (60 min limit)
Returning best findings based on investigation so far:
[Partial findings]
What's next?
1. Use current findings - Proceed with partial answer
2. Retry with extended timeout - Continue research (80 min)
3. Other
No solution found:
Research exhausted (Level 3 complete, no definitive solution)
Attempted:
✓ Local troubleshooting docs (0 matches)
✓ Rack SDK documentation (API exists but undocumented)
✓ VCV Community + GitHub (no clear consensus)
✓ Parallel investigation (3 approaches, all have significant drawbacks)
Recommendations:
1. Post to VCV Community with detailed investigation
2. Try experimental approach: [Suggestion]
3. Consider alternative feature: [Workaround]
I can help formulate community post if needed.
Subagent failure (Level 3):
⚠️ Subagent [N] failed to complete research
Proceeding with findings from other subagents (N-1 completed)...
Continue with available findings, note partial investigation in report.
WebSearch unavailable: Level 2 proceeds with SDK docs only:
⚠️ Web search unavailable
Proceeding with Rack SDK documentation only.
Results may be incomplete for community knowledge.
When executing this skill:
Common pitfalls:
Level 1:
Level 2:
Level 3:
Overall:
Common research topics:
CV Standards:
DSP Patterns:
Rack SDK APIs:
Panel Design:
Performance:
Build System: