BumpyClock/agent-templates / skills_archive/dispatching-parallel-agents
dispatching-parallel-agents
UnverifiedUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies
Dispatching Parallel Agents
Use this to investigate multiple independent failures or domains in parallel.
[codex-specific-instructions]: When using Codex, confirm with the user if they want you to proceed with subagent-driven-development before starting. Without explicit user ask, do the work yourself.
Core principle: One agent per independent domain.
Rules (Tight)
- Use only when failures/tasks are independent and do not require shared state or sequential ordering.
- If you already have a plan, use
subagent-driven-developmentfor execution and reviews. - If failures might be related or require full-system context, investigate sequentially.
- One agent per domain; prompts must be focused, self-contained, and explicit about constraints and expected output.
- Agents must follow
programmingguidance and apply TDD when required by the programming skill or explicit user request. - If investigations lead to fixes, convert them into tasks and execute via
subagent-driven-developmentso review gates apply. - Parallelize only when file scopes do not overlap.
Workflow
- Group failures by domain (test file/subsystem).
- Write one prompt per domain with context (errors, files, constraints, expected output).
- Dispatch agents in parallel.
- Review outputs, resolve conflicts, run full tests, integrate changes.
Prompt Template (Minimal)
- Scope: single file or subsystem
- Goal: what to fix or explain
- Constraints: what not to change
- Context: error text, failing tests, relevant files
- Output: summary + proposed fixes
Common Mistakes
- Too broad scope -> agent gets lost
- Missing context -> wrong fix
- No constraints -> unnecessary refactor
- Overlapping file scopes -> conflicts
Verification
- Read each summary and verify claims against code.
- Run the full test suite after integrating all changes.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 464k | Updated | Claude CodeCodex | Verified | Install |
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 464k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 318k | Updated | Claude CodeCodex | Unverified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 318k | Updated | Claude CodeCodex | Unverified | Install |