Concurrent investigation pattern - dispatches multiple AI agents to investigate and fix independent problems simultaneously.
Dispatches multiple AI agents to investigate and fix independent problems simultaneously.
npx claudepluginhub lerianstudio/ringThis skill inherits all available tools. When active, it can use any tool Claude has access to.
When you have multiple unrelated failures (different test files, different subsystems, different bugs), investigating them sequentially wastes time. Each investigation is independent and can happen in parallel.
Core principle: Dispatch one agent per independent problem domain. Let them work concurrently.
Decision flow: Multiple failures? → Are they independent? (No → single agent) | Independent? → Can work in parallel? (No/shared state → sequential) | Yes → Parallel dispatch
Use when: 3+ test files with different root causes | Multiple subsystems broken independently | Each problem understood without others | No shared state
Don't use when: Failures related (fix one might fix others) | Need full system state | Agents would interfere
1. Identify Independent Domains: Group failures by what's broken (File A: approval flow, File B: batch behavior, File C: abort). Each domain independent.
2. Create Focused Agent Tasks: Each agent gets: specific scope (one file/subsystem), clear goal (make tests pass), constraints (don't change other code), expected output (summary of findings/fixes).
3. Dispatch in Parallel: Task("Fix agent-tool-abort.test.ts") + Task("Fix batch-completion.test.ts") + Task("Fix tool-approval-races.test.ts") - all concurrent.
4. Review and Integrate: Read summaries → verify no conflicts → run full test suite → integrate all changes.
Good prompts are: Focused (one problem domain), Self-contained (all context included), Specific output (what to return).
Example: "Fix 3 failing tests in agent-tool-abort.test.ts: [list tests + expected behavior]. Timing/race issues. Read tests → identify root cause → fix (event-based waiting, not timeout increases). Return: Summary of findings and fixes."
| ❌ Bad | ✅ Good |
|---|---|
| Too broad: "Fix all tests" | Specific: "Fix agent-tool-abort.test.ts" |
| No context: "Fix race condition" | Context: Paste error messages + test names |
| No constraints: Agent refactors everything | Constraints: "Do NOT change production code" |
| Vague output: "Fix it" | Specific: "Return summary of root cause and changes" |
Related failures (fix one might fix others) | Need full context | Exploratory debugging | Shared state (same files/resources)
Scenario: 6 failures across 3 files after refactoring. Decision: Independent domains → parallel dispatch. Results: Agent 1 (timeouts → events), Agent 2 (event structure bug), Agent 3 (async wait). All independent, no conflicts, suite green. 3 problems solved in time of 1.
Parallelization (simultaneous) | Focus (narrow scope) | Independence (no interference) | Speed (3 → 1 time unit)
After agents return: Review summaries → check for conflicts → run full suite → spot check for systematic errors.
STOP and report if:
| Decision Type | Blocker Condition | Required Action |
|---|---|---|
| Independence check | Problems share state or files | STOP and use sequential investigation instead |
| Minimum threshold | Fewer than 3 independent failures | STOP and investigate directly without parallel dispatch |
| Conflict detection | Agent changes would overlap | STOP and re-partition domains |
| Context requirements | Problem requires full system context | STOP and use single agent investigation |
The following requirements CANNOT be waived:
| Severity | Condition | Required Action |
|---|---|---|
| CRITICAL | Dispatched agents for related/connected failures | MUST stop and switch to sequential investigation |
| CRITICAL | Agent changes conflict with each other | MUST re-partition domains and re-dispatch |
| HIGH | Agent prompt lacks scope constraints | MUST add specific "do NOT change" constraints |
| HIGH | Skipped post-integration test suite | MUST run full test suite before completion |
| MEDIUM | Agent prompt lacks expected output format | Should add specific output requirements |
| LOW | Could optimize domain partitioning | Fix in next iteration |
| User Says | Your Response |
|---|---|
| "Just dispatch agents for all failures, don't analyze" | "MUST verify independence first. Related failures need sequential investigation to avoid conflicts." |
| "Two failures is enough for parallel dispatch" | "CANNOT dispatch for fewer than 3 failures. Direct investigation is faster for 1-2 problems." |
| "Skip the verification step, agents finished successfully" | "MUST run full test suite after integration. Agent success doesn't guarantee no conflicts." |
| "Give agents broad scope to fix more issues" | "Focused scope prevents interference. MUST limit each agent to one file/subsystem." |
| Rationalization | Why It's WRONG | Required Action |
|---|---|---|
| "These failures look independent enough" | Looking independent ≠ verified independent; check for shared state | MUST verify independence before dispatch |
| "Dispatching for 2 failures saves time" | Overhead of parallel dispatch exceeds benefit for <3 failures | MUST investigate directly if fewer than 3 |
| "Agents can figure out their own scope" | Unclear scope causes agents to interfere with each other | MUST provide explicit scope and constraints |
| "Test suite passed for each agent" | Individual success doesn't catch integration conflicts | MUST run full suite after integration |
| "Re-partitioning is too much work" | Conflicting changes create more work than re-partitioning | MUST re-partition if domains overlap |
Creating algorithmic art using p5.js with seeded randomness and interactive parameter exploration. Use this when users request creating art using code, generative art, algorithmic art, flow fields, or particle systems. Create original algorithmic art rather than copying existing artists' work to avoid copyright violations.
Applies Anthropic's official brand colors and typography to any sort of artifact that may benefit from having Anthropic's look-and-feel. Use it when brand colors or style guidelines, visual formatting, or company design standards apply.
Create beautiful visual art in .png and .pdf documents using design philosophy. You should use this skill when the user asks to create a poster, piece of art, design, or other static piece. Create original visual designs, never copying existing artists' work to avoid copyright violations.