metaswarm

Plugin metaswarm in dsifry/metaswarm. Commands read from the repo's manifest — yesterday.

Skills (14)

• brainstorming-extension

  Loads the assumption ledger and misreadings checklist at the start of brainstorming, and enforces the design review gate when a project-sized design doc is committed — bridges superpowers:brainstorming into the metaswarm quality pipeline

• create-issue

  Create comprehensive GitHub issues with TDD plans, acceptance criteria, and agent instructions for autonomous PR lifecycle management

• design-review-gate

  Automatic review gate that runs when brainstorming commits a project-sized design doc - spawns PM, Architect, Designer, Security, UX Reviewer and CTO agents in parallel, iterates under the severity rule until all approve or it escalates

• external-tools

  Delegate implementation and review tasks to external AI CLI tools (Codex, Gemini) with cross-model adversarial review

• handling-pr-comments

  Address PR review feedback systematically — fetch inline comments, review bodies, handle outside-diff-range comments, resolve threads with proper attribution

• handoff

  Analyze the current session and write a self-contained handoff document so a fresh agent can resume the work with full context — outputs a single "Read XXX.md and do YYY." sentence

• migrate

  Migrate from npm-installed metaswarm to the marketplace plugin — removes redundant files with safety checks

• orchestrated-execution

  4-phase execution loop for work units - IMPLEMENT, VALIDATE, ADVERSARIAL REVIEW, COMMIT

• plan-review-gate

  Automatic adversarial review gate that spawns 3 independent reviewers in parallel once a plan is written - all must PASS before writing-plans presents the plan

• pr-shepherd

  Monitor a PR through to merge — handle CI failures, review comments, and thread resolution automatically until all checks pass

• setup

  Interactive project setup — detects your project, configures metaswarm, writes project-local files

• start

  Use when starting work on any task, when the user mentions metaswarm, or when the user wants to begin tracked development work

• status

  Diagnostic status report — shows metaswarm installation state, project setup, and potential issues

• visual-review

  Take screenshots of web pages and UI using Playwright for visual review and iteration

Slash commands (16)

• /metaswarm:brainstorm

  Refine an idea into a design with brainstorming, then run the design review gate for project-sized designs

• /metaswarm:create-issue

  Create a well-structured GitHub issue with acceptance criteria and agent instructions

• /metaswarm:external-tools-health

  Check the status of external AI tools (Codex CLI, Gemini CLI) and their configuration

• /metaswarm:handle-pr-comments

  Handle PR review comments: verify each finding, fix or reply, and resolve the threads

• /metaswarm:handoff

  Analyze the session and write a self-contained handoff document so a fresh agent can resume the work — outputs a single "Read XXX.md and do YYY." sentence

• /metaswarm:metaswarm-setup

  Deprecated alias for /setup

• /metaswarm:metaswarm-update-version

  Deprecated alias for /update

• /metaswarm:pr-shepherd

  Monitor a PR through to merge - handles CI failures, reviews, and thread resolution

• /metaswarm:prime

  Load relevant knowledge base facts into context before starting work

• /metaswarm:review-design

  Trigger the parallel design review gate with 6 specialist reviewers

• /metaswarm:self-reflect

  Extract learnings from recent PR reviews, conversations, and session patterns to update the knowledge base

• /metaswarm:setup

  Interactive project setup: detect the project, configure metaswarm and write project-local files

• /metaswarm:start-task

  Begin tracked work on a task with complexity assessment

• /metaswarm:start

  Alias for /start-task: begin tracked work on a task

• /metaswarm:status

  Show metaswarm diagnostic information

• /metaswarm:update

  Check for and apply metaswarm updates, then regenerate the project's phase agents

Subagents (23)

• adversarial-reviewer

  Fresh-context adversarial review of an implementation against its spec and rubrics. Read-only; use for the ADVERSARIAL REVIEW phase.

• architect-agent

  Implementation planning and architecture design. Use when a plan must be drafted from research findings.

• coder

  TDD implementation of features and fixes. Use for the IMPLEMENT phase of a work unit.

• cto-agent

  Plan review and architectural guidance. Use to review plans against codebase standards and the plan-review rubric.

• customer-service-agent

  User issue investigation and support. Use to investigate customer-reported problems with read-only access to billing, analytics and data.

• design-reviewer

  Design review gate reviewer. The role (PM, Architect, Designer, Security, UX or CTO) is passed in the prompt. Read-only.

• designer-agent

  UX, API design and developer-experience review. Use as the Designer role in the design review gate.

• issue-orchestrator

  Main coordinator for a single GitHub Issue lifecycle. Use to drive one issue from research through PR.

• knowledge-curator-agent

  Knowledge extraction and curation into the BEADS knowledge base. Use after a PR merge or to reflect on completed work.

• metrics-agent

  Collects, aggregates and reports on agent swarm performance. Use for scheduled or manual metrics reports.

• plan-reviewer

  Plan review gate adversarial reviewer. The role (Feasibility, Completeness or Scope & Alignment) is passed in the prompt. Read-only.

• planner

  Plan and decompose: turns research findings into an implementation plan and work units. May write plan files.

• pr-shepherd-agent

  PR lifecycle management through to merge. Use to monitor CI, review comments and thread resolution on a PR.

• product-manager-agent

  Use case validation and user-benefit review. Use as the Product Manager role in the design review gate.

• release-engineer-agent

  Safe delivery of approved code from merge through production verification. Use for release and deploy workflows.

• researcher

  Codebase exploration and prior-art research before planning. Read-only.

• security-auditor-agent

  Security vulnerability detection and OWASP compliance review of implemented code. Use for security audits.

• security-design-agent

  Security review of designs before implementation. Use as the Security role in the design review gate.

• slack-coordinator-agent

  Human-agent communication bridge via Slack. Use to route notifications and human responses.

• sre-agent

  Production system monitoring and incident response. Use to investigate production incidents.

• swarm-coordinator-agent

  Meta-orchestrator for multiple GitHub Issues and BEADS epics in parallel. Use to coordinate a swarm across worktrees.

• test-automator-agent

  Test writing and coverage analysis. Use to add or improve tests and check coverage.

• ux-reviewer-agent

  UX review of user flows, UI states and integration wiring. Use as the UX role in the design review gate.

Hooks (2)

• SessionStart

  "${CLAUDE_PLUGIN_ROOT}/hooks/session-start.sh"

• PreCompact

  "${CLAUDE_PLUGIN_ROOT}/hooks/session-start.sh"

Similar plugins