Dev & Engineering parallel-agentssubagent-orchestrationdebuggingtest-failuressoftware-development

Parallel Agent Dispatch

Delegate independent engineering problems to focused agents and investigate them concurrently.

FollowSkills review · FSRS-2.0
Use with care
50/ 100 5-point scale 2.5 / 5
1 2 3 4 5 6
1Trust14 / 25 · 2.8/5

The document scopes each agent to an independent problem domain, supplies explicit constraints, and requires review for conflicts and full-suite verification. It does not specify user confirmation, agent permission boundaries, sensitive-data handling, data-flow disclosure, rollback, or controls for external side effects, so points are deducted.

2Reliability8 / 20 · 2.0/5

Trigger conditions, parallel-dispatch patterns, prompt structure, and post-dispatch integration checks are clear, including a warning against merely increasing timeouts. However, the static material does not establish executable reproduction and lacks diagnosable handling for dispatch failure, shared-resource conflicts, malformed input, or unresponsive agents; the static cap and uncertainty therefore limit the score.

3Adaptability10 / 15 · 3.3/5

The skill clearly targets multiple independent problems that can be worked in parallel and identifies related failures, shared state, and exploratory debugging as non-fit cases. Input/output boundaries, concrete environment support, Chinese-language support, and mainland-China network fit are not documented, so points are deducted.

4Convention8 / 15 · 2.7/5

The skill is well layered, with overview, use/non-use conditions, workflow, prompt examples, common mistakes, limitations, and verification guidance. Repository context supplies an MIT license, version, and maintenance/community entry points, but the skill itself lacks installation notes, version history, changelog, and explicit maintenance responsibility, so points are deducted.

5Effectiveness6 / 15 · 2.0/5

The method provides directly usable task decomposition, agent prompts, and integration steps for independent failure investigations, with a defined summary output. Claims of time savings, zero conflicts, and a green full suite are only document assertions under static review; human review and full testing remain necessary, so points are deducted.

6Verifiability4 / 10 · 2.0/5

The skill includes concrete procedures, a decision diagram, prompt examples, and a verification checklist, while repository context indicates test infrastructure. It lacks skill-specific reproducible tests or independent corroboration, and the real-session example is not independently substantiated, so points are deducted.

Evidence confidence:Low Reviewed Jul 19, 2026 Reviewed revision d884ae04edeb
The upstream repository has new commits since this review. The score still applies to the reviewed revision shown and may not cover the latest changes.
Before you use it
  • Parallel agents may still modify shared code or resources; independence should be confirmed before dispatch and conflicts and regressions checked before integration.
  • The document does not define agent permissions, sensitive-information minimization, user confirmation, or failure rollback; do not assume agents may safely perform arbitrary changes.
  • The claimed real-world example and impact figures were not independently verified in this static review.
Review evidence [1][2][3][4][5][6][7]
See the full review method →

What does this skill do, and when should you use it?

Parallel Agent Dispatch is one skill in obra/superpowers for handling two or more independent tasks without shared state or sequential dependencies. It assigns each problem domain a focused, self-contained agent prompt and dispatches the agents in parallel. After they return, the coordinating agent reviews their summaries, checks for conflicts, runs the full test suite, and integrates the results. It is not intended for related failures, exploratory debugging, full-system investigations, or work involving shared resources.

Identifies independent problem domains; creates agent prompts with a specific scope, goal, constraints, and expected output; dispatches multiple subagents in the same response; reviews returned summaries; checks for conflicting edits; runs the full test suite; integrates and spot-checks the results.

  1. A developer has failures in several test files with different root causes and wants one investigation per file or subsystem.
  2. An engineering team is debugging independent issues across separate subsystems.
  3. Each investigation can proceed without context from the others and parallel work can reduce total investigation time.
  4. A developer wants agents to stay within explicit file or subsystem boundaries and return a clear root-cause summary.

What are this skill's strengths and limitations?

Pros
  • Provides focused, self-contained context for each problem domain.
  • Allows independent investigations to run concurrently.
  • Defines scope constraints and expected summaries.
  • Includes a review, conflict-checking, integration, and verification workflow.
Limitations
  • Unsuitable for related failures, exploratory debugging, or tasks requiring full-system context.
  • Shared files, resources, or state can cause agent interference.
  • The source does not provide a standalone test suite or release history for this skill.
  • The coordinator still must review changes and run the full test suite after dispatch.

How do you install this skill?

This skill is installed as part of the Superpowers collection; the supplied material does not document a standalone installation. In Claude Code, run /plugin install superpowers@claude-plugins-official. In Codex CLI, run /plugins, search for superpowers, and select Install Plugin. Other supported harnesses have separate collection installation instructions in the README.

How do you use this skill?

Give the coding agent a task containing independent problems, for example: Investigate these three independent test failures in parallel: agent-tool-abort.test.ts, batch-completion-behavior.test.ts, and tool-approval-race-conditions.test.ts. Give each problem its own focused subagent, avoid changing unrelated code, and return a summary of the root cause and fix. Use it only when the problems can safely be worked on concurrently.

FAQ

Does the skill automatically run agents in parallel?
It instructs the host agent to issue multiple subagent dispatches in the same response; the supplied material does not specify the host's underlying scheduling implementation.
When should I avoid using it?
Avoid it when failures are related, tasks depend on shared state, a full-system investigation is required, or agents could interfere by editing the same resources.
What permissions or dependencies does it require?
The skill names no external service, network dependency, or package. The work does require agents to read and modify project files and run the project's tests.

More skills from this repository

All from obra/superpowers

Dev & Engineering

Systematic Debugging

Find the root cause before changing code, reducing trial-and-error rework.

Dev & Engineering

Superpowers Skill Bootstrap

Ensures applicable skills are discovered and invoked before every response or action.

Dev & Engineering

Executing Plans

Turn a written implementation plan into verified, checkpointed development work.

Dev & Engineering

Continuous Code Review

Dispatch focused reviews after tasks, major features, and before merges to catch problems early.

Dev & Engineering

Writing Plans

Turn multi-step software requirements into executable, testable implementation plans.

Dev & Engineering

Design Before Code

Clarify intent, compare approaches, and secure an approved design before implementation begins.

Dev & Engineering

Subagent-Driven Development

Execute independent implementation tasks with fresh agents and review gates.

Dev & Engineering

Test-Driven Skill Authoring

Create, revise, and verify Agent Skills with a test-driven documentation workflow.

Dev & Engineering

Test-Driven Development

Prove expected behavior with a failing test before writing implementation code.

Dev & Engineering

Development Branch Finisher

Safely close completed development work after tests pass, with guided merge, PR, retention, or discard choices.

Dev & Engineering

Rigorous Code Review Reception

Verify review feedback before changing code, then act on technical evidence.

Dev & Engineering

Isolated Git Workspace

Creates an isolated workspace for feature work, then prepares the project and verifies its clean test baseline.

Dev & Engineering

Verification Before Completion

Require fresh evidence before claiming work is complete, fixed, or passing.

Related skills