task-management-agents

Alembic migration workflow for SQLAlchemy model changes, emphasizing autogeneration only, date-prefixed migration filenames, env.py model imports, review of generated revisions, and validation with upgrade commands and integration tests.

[NOT APPLICABLE — Ezra-specific reference; do not invoke in this project] Backend architecture patterns, API design, database optimization, and server-side best practices for Node.js, Express, and Next.js API routes.

[NOT APPLICABLE — Ezra-specific reference; do not invoke in this project] Clerk authentication patterns for the Ezra Next.js monorepo — server vs client auth, middleware strategies, Server Actions, API route protection, caching with auth, and token injection via authTokenProvider for Orval-generated hooks. Use when adding Clerk auth to a new app, protecting routes or actions, wiring auth tokens into API calls, debugging "unauthorized" or 401/403 errors, asking "how does auth work in this app", checking why an API call returns 401, protecting a page or Server Action, or implementing role/permission checks. Also use when touching middleware.ts, ClerkProvider, ClerkTokenSync, authTokenProvider, or any import from @clerk/nextjs. Do NOT use for Clerk SDK installation or initial project setup (see Clerk docs quickstart).

[NOT APPLICABLE — Ezra-specific reference; do not invoke in this project] Universal coding standards, best practices, and patterns for TypeScript, JavaScript, React, and Node.js development.

Advisory skill for evaluating and organizing backend domain packages according to DDD best practices and capability-oriented design. Use when creating new domains, reviewing existing domain boundaries, refactoring packages, or deciding where new business logic should live.

[NOT APPLICABLE — Ezra-specific reference; do not invoke in this project] Sentry SDK internals for the ezra_observability Sentry adapter — sentry_sdk.init configuration, StarletteIntegration/FastApiIntegration/SqlalchemyIntegration setup, EventScrubber denylist, before_send filters, PII rules, and adapter-level tests. ADAPTER AUTHORS ONLY. App code must use `from ezra_observability import get_observability` (see ezra-add-monitoring).

[NOT APPLICABLE — Ezra-specific reference; do not invoke in this project] Sentry SDK internals for the @ezra/observability Next.js Sentry adapter — Sentry.init configuration across client/server/edge runtimes, withSentryConfig, session replay setup, beforeSend filters, and adapter-level tests. ADAPTER AUTHORS ONLY. App code must use `import { getObservability } from '@ezra/observability'` (frontend skill TBD; backend mirror is ezra-add-monitoring).

Universal error tracking patterns — error capture, context enrichment, breadcrumbs, PII scrubbing, before-send filtering, performance tracing, and alerting hygiene

[NOT APPLICABLE — Ezra-specific reference; do not invoke in this project] Add error tracking to Ezra backend code via the ezra_observability port. Use when adding a new FastAPI route, a new Temporal activity, a new background task, a new external API integration, or any code path that needs error capture, user context, or breadcrumbs. Triggers on "add monitoring", "add sentry", "capture this error", "track this", "instrument this endpoint", "report this exception", "add error tracking". Do NOT use for tracing/profiling work (errors-only scope today) or for Sentry adapter internals — see error-tracking-fastapi-sentry for adapter author work.

[NOT APPLICABLE — Ezra-specific reference; do not invoke in this project] Analyze and address unresolved review comments for pull requests with thoughtful responses and code fixes

[NOT APPLICABLE — Ezra-specific reference; do not invoke in this project] Analyzes codebase implementation details. Call the codebase-analyzer agent when you need to find detailed information about specific components. As always, the more detailed your request prompt, the better! :)

[NOT APPLICABLE — Ezra-specific reference; do not invoke in this project] Locates files, directories, and components relevant to a feature or task. Call `codebase-locator` with human language prompt describing what you're looking for. Basically a "Super Grep/Glob/LS tool" — Use it if you find yourself desiring to use one of these tools more than once.

[NOT APPLICABLE — Ezra-specific reference; do not invoke in this project] codebase-pattern-finder is a useful subagent_type for finding similar implementations, usage examples, or existing patterns that can be modeled after. It will give you concrete code examples based on what you're looking for! It's sorta like codebase-locator, but it will not only tell you the location of files, it will also give you code details!

[NOT APPLICABLE — Ezra-specific reference; do not invoke in this project] Check active Linear tickets with comprehensive status overview

[NOT APPLICABLE — Ezra-specific reference; do not invoke in this project] Create git commits and relevant branches with user approval and no Claude attribution

[NOT APPLICABLE — Ezra-specific reference; do not invoke in this project] Generate WhatsApp community content for the Ezra referral network. Use when creating deal pitches, 'looking for deals' posts, 'looking for introductions' posts, or any community announcement. Triggers on: 'community post', 'deal pitch', 'looking for deals', 'write announcement', 'referral community', 'WhatsApp announcement', 'community manager'. Do NOT use for general marketing tasks unrelated to the WhatsApp community — use the individual marketing skills (copywriting, social-content, marketing-psychology, referral-program) instead.

[NOT APPLICABLE — Ezra-specific reference; do not invoke in this project] Create handoff document for transferring work to another session

[NOT APPLICABLE — Ezra-specific reference; do not invoke in this project] Create well-structured Linear tickets via MCP following Ezra's hierarchy (Initiative > Project > Issue > Sub-Issue). Use when creating tickets, filing bugs, breaking down features, or converting Slack threads into Linear issues. Triggers on "create ticket", "file issue", "linear ticket", "track this".

[NOT APPLICABLE — Ezra-specific reference; do not invoke in this project] Create detailed implementation plans through interactive research and iteration

[NOT APPLICABLE — Ezra-specific reference; do not invoke in this project] Orchestrate full PR preparation — code review, lint/format/test with auto-healing, documentation review, and PR creation via gh CLI. Use when dev says "prepare PR", "create PR", "submit PR", "PR ready", "open PR", "push PR", "ready for review", "ship it", "send PR", "merge request". Do NOT use for just reviewing code (use ezra-review-code) or just committing (use ezra-commit).

[NOT APPLICABLE — Ezra-specific reference; do not invoke in this project] Generate PR descriptions from branch changes for creating or updating pull requests

[NOT APPLICABLE — Ezra-specific reference; do not invoke in this project] Generate pytest tests for Python backend code with planning, generation, and automated healing

[NOT APPLICABLE — Ezra-specific reference; do not invoke in this project] Generate e2e Playwright tests from test plan specs with automated healing and refinement

[NOT APPLICABLE — Ezra-specific reference; do not invoke in this project] Generate Vitest unit and integration tests for frontend packages with automated healing and refinement

[NOT APPLICABLE — Ezra-specific reference; do not invoke in this project] GitHub PR operations via gh CLI. Use when creating reviews, replying to comments, or fetching PR data. CRITICAL - Always create reviews as drafts by omitting the event field.

[NOT APPLICABLE — Ezra-specific reference; do not invoke in this project] Implement technical plans from docs/plans/ with verification

[NOT APPLICABLE — Ezra-specific reference; do not invoke in this project] Iterate on existing implementation plans with thorough research and updates

[NOT APPLICABLE — Ezra-specific reference; do not invoke in this project] Use when the task is "create a Langfuse prompt", "list Langfuse prompts", "version this prompt in Langfuse", or "connect a Langfuse-managed prompt to a PydanticAI agent". Focuses on prompt management through the Langfuse Python SDK and minimal integration with `ezra_llm`.

[NOT APPLICABLE — Ezra-specific reference; do not invoke in this project] Use when the task is “instrument this PydanticAI agent”, “add Langfuse tracing”, “add Langfuse evals”, “make this agent observable”, “fetch prompts from Langfuse”, “manage system prompts”, “version prompts”, or “wire Langfuse prompt into agent”. Covers observability, prompt management, and eval capture via `ezra_llm`.

[NOT APPLICABLE — Ezra-specific reference; do not invoke in this project] Set up worktree for reviewing colleague's branch

[NOT APPLICABLE — Ezra-specific reference; do not invoke in this project] Document codebase as-is with comprehensive research

[NOT APPLICABLE — Ezra-specific reference; do not invoke in this project] Resume work from handoff document with context analysis and validation

[NOT APPLICABLE — Ezra-specific reference; do not invoke in this project] Perform FAANG-level code review analyzing necessity, correctness, and completeness of changes

[NOT APPLICABLE — Ezra-specific reference; do not invoke in this project] Rigorous documentation reviewer for "fresh eyes" passes. Finds concrete problems, checks template compliance, and makes targeted fixes.

[NOT APPLICABLE — Ezra-specific reference; do not invoke in this project] Automated retrospective that reviews session history to identify improvements.

[NOT APPLICABLE — Ezra-specific reference; do not invoke in this project] Scaffold a new frontend or backend app in the Ezra platform monorepo. Use when creating a new Next.js frontend app or FastAPI backend app, when user says "add app", "new app", "scaffold app", "create app", or "new service". Handles all registration points — workspace, Docker Compose, Dockerfile, CI/CD, infrastructure stubs.

[NOT APPLICABLE — Ezra-specific reference; do not invoke in this project] Scaffold pytest testing infrastructure for a new backend package. Use when creating a new app in backend/packages/apps/ or a new domain in backend/packages/domains/, when user says "add tests to", "scaffold tests for", "setup testing for" a backend package, or when a new package needs conftest.py, factories, and smoke tests. Also use when retrofitting tests onto an existing package that has no tests yet.

[NOT APPLICABLE — Ezra-specific reference; do not invoke in this project] Scaffold Vitest unit/integration + Playwright E2E testing infrastructure for a new frontend app. Use when creating a new app in frontend/apps/, when user says "add tests to", "scaffold tests for", "setup testing for" a frontend app, or when a new Next.js app needs vitest config, E2E with Testcontainers, POMs, and Playwright fixtures. Also use when retrofitting tests onto an existing app that has no tests yet.

[NOT APPLICABLE — Ezra-specific reference; do not invoke in this project] Scaffold Vitest testing infrastructure for a new shared frontend package (unit + integration, no E2E). Use when creating a new package in frontend/packages/, when user says "add tests to", "scaffold tests for", "setup testing for" a shared package, or when a new package like hooks, forms, or validators needs vitest config and smoke tests. Also use when retrofitting tests onto an existing package that has no tests yet. Do NOT use for frontend apps (use ezra-scaffold-frontend-app-tests instead).

[NOT APPLICABLE — Ezra-specific reference; do not invoke in this project] TimelinesAI API patterns for WhatsApp automation in Ezra. Use when building n8n workflows, agents, or services that interact with WhatsApp via TimelinesAI. Covers: posting to communities, reading messages, file uploads, webhooks, sub-group readability, and known limitations. Triggers on: 'TimelinesAI', 'WhatsApp API', 'WhatsApp automation', 'post to community', 'WhatsApp webhook', 'community announcement'.