co-dev

Analyze, document, map, or scan the database schema. Use when the user wants to analyze the database, document the database, generate schema docs, map the database, create DB documentation, or inspect the database structure. Generates a docs/db.md file with complete database schema documentation. Auto-detects language/framework. Supports MySQL, PostgreSQL, SQLite, MongoDB, Elasticsearch, Redis, and BigQuery.

Manage App Store Connect apps, builds, distribution, or Xcode Cloud. Use when the user wants to check builds, manage TestFlight beta groups and testers, read or respond to App Store reviews, manage in-app purchases and their price schedules, list app versions, check app status, list Xcode Cloud workflows or build runs, start an Xcode Cloud build, or investigate Xcode Cloud build failures, logs, and test results.

Manage AWS infrastructure, services, or resources. Use when the user wants to manage any AWS service including EC2, S3, Lambda, RDS, DynamoDB, ECS, EKS, IAM, CloudWatch, CloudFront, Route53, SQS, SNS, API Gateway, CloudFormation, Secrets Manager, VPC, ELB, or any other AWS service. Also use for searching AWS documentation, checking costs, or running any AWS CLI command.

Exhaustive multi-phase code audit using parallel agents (security, dependencies, code quality, infrastructure, tests, and more). Use when the user wants a deep or thorough code review, a comprehensive audit, a security-and-quality sweep of a repo or subsystem, or a multi-agent review that goes beyond the current diff. Optionally scoped to a path or subsystem.

House standards for authoring source code and tests — above all, comment discipline (default none, hard cap one line) and the diff sweep that enforces it. Use whenever writing, editing, porting or reviewing source code or test files, including a one-line fix or a single added test, and before committing or opening a PR. Also use when deciding whether a comment earns its place, where design rationale belongs, or why a comment block was deleted.

Query, list, analyze, or investigate Firebase Crashlytics crash data. Use when the user wants to check crashes, list top crashes, investigate a crash, get stack traces, view crash trends, analyze crash data, or find crash issues. Queries Crashlytics data exported to BigQuery via the bq CLI. Supports Android, iOS, and tvOS.

Create, open, file, or report an issue, bug, or ticket in GitHub or Jira. Use when the user wants to open an issue, file a bug, report a bug, create a ticket, log an issue, or submit a bug report. Automatically detects if GitHub issues are enabled; if so creates a GitHub issue, otherwise creates a Jira issue.

Create, open, submit, or prepare a pull request (PR). Generates the commit message, PR title, and PR body, opens the PR, then returns the repo to its default branch. Use when the user wants to create a PR, open a PR, submit a PR, make a PR, push a PR, send a PR, generate PR content, prepare a pull request, or fill a PR template from code changes.

Verify that every item (and its sub-tasks) listed in a section of a Google Doc has a matching equivalent in monday.com and/or Jira, per team. Reads a planning/roadmap doc as the source of truth, resolves each team to its tracker(s) (Monday board(s) and/or Jira project(s), confirmed once and cached), fuzzy-matches doc items to tracker items, and writes a read-only coverage report listing matched / missing / ambiguous items with a best-effort sub-task ratio. Use when the user wants to check a Google Doc against Monday/Jira, confirm a roadmap or plan is tracked, audit tracker coverage of a doc, find untracked items, reconcile a planning doc with the trackers, or verify items exist in Monday and/or Jira for each team. The downstream weekly-dev-report (Jira) and monday-weekly-report (Monday) then track the matched items. Writes COVERAGE_REPORT.md and prints to stdout.

Manage Google Cloud infrastructure, services, or resources. Use when the user wants to manage any GCP service including Compute Engine, Cloud Run, Cloud SQL, GKE, Cloud Storage, IAM, Pub/Sub, Cloud Logging, Cloud Monitoring, Firebase, Artifact Registry, Secret Manager, Cloud Functions, App Engine, Dataflow, BigTable, Spanner, Cloud CDN, Cloud DNS, VPC, Load Balancing, or any other Google Cloud service. Also use for running any gcloud, gsutil, or kubectl command.

Manage Heroku applications, deployments, or infrastructure. Use when the user wants to check a Heroku app, view logs, scale dynos, restart, manage Postgres databases, view config vars, manage addons, check pipelines, enable maintenance mode, manage domains, view metrics, or check app health.

Manage Loco (localise.biz) translation assets. Use when the user wants to create a translation key, delete a translation key, translate text, manage localization, add a Loco asset, remove unused translations, scan for unused tokens, or manage i18n keys. Supports create, delete, and scan commands with multi-project support and auto-translation.

Migrate, port, translate, rewrite, or convert a codebase from one language, framework, or runtime to another (e.g. Python→TypeScript, JavaScript→TypeScript, Zig/C++→Rust, Java→Kotlin, Angular→React, Express→Fastify, Rails→Phoenix, Enzyme→React Testing Library, CommonJS→ESM). Use when the user wants to migrate code, port a project, translate a language, rewrite a service in another stack, do a large-scale mechanical refactor, or run a framework/version upgrade that spans many files. Runs a six-step engine: build a rulebook, map dependencies, inventory gaps, stress-test, then translate → compile → test → verify with parallel agents and adversarial review.

Manage monday.com boards, items, and work management data. Use when the user wants to look up or create items, search a board, update item status or column values, add an update/comment, move items between groups, create boards, groups, or columns, inspect a board's schema, look up users or teams, or create and read WorkForms in monday.com.

Generate a weekly project-status report from one or more monday.com boards — general project tracking, not engineering throughput. Covers per-board health, owner activity (role-aware), items moved this week, overdue and due-soon work, stuck/blocked and stale items, upcoming milestones, and a date-driven delivery-risk assessment with a concrete catch-up plan. Role-aware (three roles — owner / external / other, confirmed once and cached) and runs against a past-full-week or week-to-date window so it can be run any day. Use when the user wants a weekly project report, project status report, monday board status, project health check, monday.com board delivery-risk assessment, or stakeholder update from monday.com. Writes PROJECT_REPORT.md and prints to stdout; emails on --send.

Query New Relic observability data or manage monitoring. Use when the user wants to run NRQL queries, check alerts, view incidents, analyze logs, get golden metrics, inspect entity health, check dashboards, view SLIs/SLOs, track deployments, monitor synthetics, check error rates, view throughput, or query any New Relic data.

Manage PayPal payments, invoices, or financial data. Use when the user wants to check payments, orders, refunds, create or view invoices, manage disputes, subscriptions, catalog products, shipment tracking, or look up any PayPal transaction data.

Fetch, filter, or analyze Google Play Store reviews, or post replies. Use when the user wants to check Google Play reviews, filter reviews by rating or date, analyze review sentiment, respond to reviews, or get review analytics for an Android app.

Query the database, run a query, look up data, search the database, or check data. Use when the user wants to query the database, run a SQL query, look up data, find data, search for records, check the database, or ask questions about data. Executes queries via CLI commands using natural language. Reads schema context from docs/db.md. Supports MySQL, PostgreSQL, SQLite, MongoDB, Elasticsearch, Redis, and BigQuery.

Review, create, update, check, write, document, or audit architecture documentation (docs/architecture.md). Use when the user wants to review the architecture, check architecture docs, write architecture docs, document the architecture, or update architecture documentation to match organizational standards with accurate technical content.

Audit AWS spend for cost optimization over the last 3 complete months, compared against the same 3 months of the previous year, using parallel agents per cost domain. Use when the user wants an AWS cost review, cloud cost optimization, a spend audit, to find AWS waste or savings, to know why the AWS bill went up, a year-over-year cost comparison, or a FinOps review. Never recommends Reserved Instances or Savings Plans. Prints the report to the terminal and waits — writes no file unless asked.

Review, audit, or improve user-facing copy — UI microcopy (error messages, empty states, confirmation dialogs, loading states, buttons/CTAs), form labels and help text, voice/tone consistency, i18n readiness, and marketing copy. Use when the user wants a copy review, to check or fix UI text, audit microcopy, improve error messages or empty states, tighten button/CTA wording, review form help text, check inclusive/plain language, find hardcoded strings, or review landing-page/marketing copy. Audits a codebase (optionally a path or the current diff) and writes docs/copy-review.md.

Review, compare, audit, or check UI code against Figma designs. Use when the user wants to compare code to a Figma design, check design implementation, audit UI fidelity, verify design compliance, or review design-to-code accuracy. Supports Android (Jetpack Compose, XML layouts), iOS (SwiftUI, UIKit), and web (HTML/CSS, React, Vue, Angular) platforms.

Review, create, update, check, fix, improve, write, or audit README.md. Use when the user wants to review the README, check the README, fix the README, write a README, improve the README, or update the README to match organizational standards with accurate project-specific content.

Review, audit, or check SEO, GEO (AI answer-engine / generative-engine optimization), and front-end web quality (performance, Core Web Vitals, accessibility). Use when the user wants an SEO audit, technical SEO check, GEO/AEO audit, AI-search optimization, llms.txt review, structured-data/schema check, meta-tag/Open-Graph audit, sitemap or robots.txt review, crawlability analysis, a Lighthouse or web-quality audit, a performance or Core Web Vitals (LCP/INP/CLS) review, or an accessibility (WCAG) audit. Audits a codebase and/or a live site URL (usually production) and writes docs/seo-audit.md.

Review, create, update, check, write, or audit the user guide (docs/user-guide.md). Use when the user wants to write a user guide, write user docs, create user documentation, review the user guide, check the user guide, update user documentation, or document the product for end users.

Run linters, lint the code, check code style, or fix linting issues. Use when the user wants to lint, run linters, check code quality, verify code style, fix linting errors, or run code checks after completing code modifications.

Summarize sprint work items grouped by repo and ~3-day blocks. Use when the user wants a sprint summary, a sprint scope report or sprint scope overview broken into work blocks, work summary, sprint breakdown, or wants to see what work is planned in a sprint. Fetches tasks and bugs from Jira (excludes stories), estimates effort from descriptions, and groups items into approximately 3-day work blocks per repository.

Audit issues in a "resolved" / "fixed" / "done" state that have not been closed yet, verify the fix actually exists in the current codebase, then either close them with a detailed summary or kick them back to the original resolver with an explanation. Use when the user wants to verify resolved issues, audit resolved tickets, audit issues marked resolved but still open, check fixed bugs, close stale resolved issues, clean up the resolved column, sweep resolved issues, sweep done tickets that are not yet closed, verify fixes, confirm resolutions, validate that fixed issues are really fixed, or review the resolved-but-not-closed backlog. Works against GitHub Issues if enabled on the repo, otherwise against Jira. Handles Jira projects with separate "Resolved" → "Closed" workflow as well as projects that use a single terminal state.

Generate a weekly team activity report from the active Jira sprint and linked GitHub repos, with per-member achievability ratings, stuck-ticket flags, stalled-member flags, and worklog audit. The roster mixes engineers, QA, content, and consultants; titles and language stay generic so non-engineering members aren't mislabelled. Flags sprint-goal delivery risk with a concrete, prioritized catch-up plan, and confirms each member's role (developer full-time / consultant part-time / manager-cto-ciso / tester / other) once interactively then caches it so later runs reuse it. Use when the user wants a weekly activity report, per-member Jira sprint progress audit, Jira sprint delivery-risk assessment, contributor status report, time-logged audit, or team check-in. Pulls roster from the active sprint, auto-discovers repos from Jira ticket dev-info, emails the report on --send, otherwise writes WEEKLY_REPORT.md and prints to stdout.

Implement one or more GitHub issues or Jira tickets end-to-end — explore, design, implement, and open a PR. Multiple issues run in parallel. Use when the user wants to work on an issue, implement a ticket, fix a bug from a tracker, or take a GitHub issue or Jira key from description to pull request.

Write, generate, or add automated tests for code — unit, integration, and end-to-end. Use when the user wants to write tests, add test coverage, generate unit tests, create specs, add a regression test, backfill missing tests, cover an untested file, or test a change before opening a PR. Auto-detects the language and the test framework already in the repo. Supports Ruby, PHP, Python, JavaScript/TypeScript, C#/.NET, Swift, Kotlin/Java, Go, C/C++, and Rust.