Devil Starswarm
Multi-perspective review via parallel Agents. Unleashes its true power on large-scale changes where cost is acceptable.
Effect
Runs Devil Chainβs 4 steps across separate Agents in parallel. Roughly 4x faster than serial execution.
| Security Agent | Perf Agent | Logic Agent | Flow Agent |
|---|---|---|---|
| β | β | β | β |
β Merge β Concern List β Fix β Rerun
When to Use
- Changes span 50+ files or cross multiple services
- Agent cost (~0.18 Γ 4 turns) is acceptable
- Speed is critical (e.g., pre-release blocker)
How It Works
1. Launch 4 Agents in Parallel
Spin up AI agents in parallel, each with a specialized prompt for Security / Performance / Logic / Flow
2. Merge Results
Combine concern lists from all Agents. Deduplicate and prioritize.
3. Fix β Rerun
If concerns remain, fix them. Large fix scope β re-Starswarm. Small scope β fall back to Chain.
Combinations
| Card | Synergy |
|---|---|
| Devil Chain | Prerequisite card. Without understanding Chain, you cannot integrate the results |
| Devil Gate | Feed the relevant checklist to each Agent (precision UP) |
Cost
Serial Chain: β 1 turn (you run it yourself)
Parallel Starswarm: β 0.72 (0.18 Γ 4 Agent turns)
Trade-off: API cost β review speed
ββ¦Surround them on all sides, and thereβs nowhere left to run.β β Eris
Want to use this card permanently? Convert it directly into a SKILL.md β Effect β Overview, When to Use β Trigger, How It Works β Steps.