compozy

Plugin compozy from marketplace novac-stack in HigorAlves/claude-plugin. Commands read from the repo's manifest — 1 month ago.

Skills (11)

• branch-completion

  Use when implementation is complete, all tests pass, and you need to decide how to integrate the work — guides completion by presenting structured options for merge, PR, or cleanup

• parallel-agents

  Use when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies

• pr-review

  Use when reviewing code — establishes comment tone, false positive guidelines, suggestion rules, and hard review rules

• qa-session

  Use when running an interactive QA session — listen to user-reported problems, reproduce them with headless Chrome, take screenshots as evidence, and file GitHub issues

• spec-authoring

  Use when writing tech specs, creating task manifests, decomposing features into tasks, or authoring architecture documents from requirements

• systematic-debugging

  Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes

• tdd

  Use when implementing any feature or bugfix, before writing implementation code

• team-agents

  Use when --team flag is passed or when a task would benefit from multiple agents collaborating rather than working independently — agents take different roles and review each other's work

• using-compozy

  Use when starting any conversation — establishes how to find and use skills, requiring Skill tool invocation before ANY response including clarifying questions

• verification

  Use when about to claim work is complete, fixed, or passing, before committing or creating PRs — requires running verification commands and confirming output before making any success claims

• worktrees

  Use when starting feature work that needs isolation from current workspace — creates isolated git worktrees with smart directory selection and safety verification

Slash commands (13)

• /compozy:code-review

  Review a GitHub PR — checks code quality, test coverage, and requirements alignment

• /compozy:debug

  Systematic debugging — find root cause before fixing, with structured investigation phases

• /compozy:design

  Brainstorming and design — explore requirements, ask questions, propose approaches, and produce a design spec

• /compozy:finish

  Complete a development branch — verify tests, present integration options, execute choice, cleanup

• /compozy:jira

  Jira ticket triage — discover, analyze, investigate the codebase, enrich the ticket with findings and acceptance criteria, and prepare for implementation

• /compozy:orchestrate

  Spec-driven development orchestration — analyze requirements, generate tech specs, decompose into parallel tasks, execute, review, and create PRs

• /compozy:plan

  Create a TDD-structured implementation plan from a design spec or requirements

• /compozy:pr-respond

  Address PR review feedback — read comments, fix code, reply to threads, push changes

• /compozy:qa

  Run an interactive QA session — reproduce issues with headless Chrome, capture screenshots, and file GitHub issues

• /compozy:resume

  Resume an interrupted Compozy orchestration from the last checkpoint

• /compozy:sentry-fix

  Fix Sentry issues — discover, analyze with rich context, find root cause, fix with TDD, verify, and resolve

• /compozy:spec

  Generate, view, or edit technical specifications without running the full orchestration pipeline

• /compozy:status

  Show pipeline status across all active Compozy orchestrations

Subagents (17)

• bug-hunter

  Scans PR diffs for real bugs, logic errors, and runtime failures. Use this agent during code review to catch issues that will break in production — syntax errors, type mismatches, incorrect logic, null reference problems, race conditions, and off-by-one errors. Focuses only on the changed code.

• code-quality-reviewer

  Stage-2 reviewer — verifies implementation is well-built (clean, tested, maintainable). Only dispatched after spec-compliance-reviewer passes.

• guidelines-reviewer

  Reviews PR changes for compliance with project guidelines (CLAUDE.md). Use this agent during code review to check that changes follow established project conventions, import patterns, error handling rules, naming conventions, and other documented standards.

• integration-validator

  Validates cross-task integration after parallel implementation — checks imports resolve, interfaces match, components wire together correctly, and the build passes. Use this agent after task execution waves complete.

• jira-analyzer

  Analyzes Jira tickets using MCP tools — gathers description, acceptance criteria, linked issues, subtasks, comments, sprint/epic context into a structured report

• jira-sync

  Creates and updates Jira tickets from compozy task manifests — syncs task decomposition to Jira for manager visibility and progress tracking

• pr-assembler

  Generates branch names, commit messages, and PR descriptions from completed implementation work. Creates the branch, commits, pushes, and opens the PR. Use this agent in the final phase of orchestration.

• prd-analyzer

  Analyzes product requirements documents, extracts structured requirements, identifies gaps and ambiguities, and generates clarifying questions. Use this agent when parsing PRDs, GitHub issues, or feature descriptions into structured requirements.

• qa-validator

  Validates acceptance criteria are met, runs regression checks, and writes missing tests following existing repo patterns. Only writes tests if the repo already has them — never introduces a new test framework.

• requirements-checker

  Checks whether a PR implements what the ticket or spec requires. Use this agent when reviewing a PR with requirements context (Jira ticket, PRD, or description) to verify coverage, identify gaps, flag partial implementations, and note scope drift.

• security-reviewer

  Reviews PR changes for security vulnerabilities and correctness issues. Use this agent during code review to catch injection attacks, auth bypasses, data exposure, unsafe deserialization, SSRF, insecure crypto, and other security problems in the introduced code.

• sentry-analyzer

  Analyzes Sentry issues using MCP tools — gathers stack traces, breadcrumbs, traces, tag distributions, and Seer AI analysis into a structured report

• spec-compliance-reviewer

  Stage-1 reviewer — verifies implementation matches the tech spec (nothing more, nothing less). Must pass BEFORE code-quality-reviewer runs.

• spec-generator

  Generates complete technical specifications from analyzed requirements and codebase context. Uses the spec-authoring skill for templates and conventions. Use this agent after PRD analysis and codebase discovery.

• task-decomposer

  Decomposes technical specifications into parallel-safe task manifests with waves, dependencies, and file ownership. Use this agent after a tech spec is approved.

• task-implementer

  Executes a single implementation task from a task manifest, creating and modifying files per the tech spec. Each instance owns exclusive files — no conflicts with parallel agents.

• test-analyzer

  Analyzes test coverage for PR changes and suggests missing test scenarios. Use this agent during code review to identify untested code paths, missing edge cases, and suggest concrete test scenarios based on the requirements context and changed code. Focuses on behavioral coverage, not line counts.

Hooks (2)

• SessionStart

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

• PreToolUse

  "${CLAUDE_PLUGIN_ROOT}/hooks/pre-commit-branch-check.sh"

Similar plugins