Skip to main content

What You Get

hatch3r ships a comprehensive agentic setup out of the box. During hatch3r init, you choose a content profile (Minimal, Standard, Full, or Custom) that determines which items are installed. Here's everything available in the Full profile.

CategoryCountHighlights
Agents16Code reviewer, test writer, security auditor, implementer (sub-agentic), fixer, researcher, architect, DevOps, and more
Skills26Bug fix, feature implementation, issue workflow, release, incident response, context health, cost tracking, recipes, customization, and more
Rules27Code standards, testing, API design, observability, theming, i18n, security patterns, agent orchestration, deep context, and more
Commands34Board init, board fill, board groom, board pickup, board refresh, planning (feature, bug, refactor, test), healthcheck, security-audit, context-health, cost-tracking, customization, and more
MCP Servers10 (3 default + 7 opt-in)Playwright, Context7, Filesystem (default); GitHub, Brave Search, Sentry, Postgres, Linear, Azure DevOps, GitLab (opt-in)

Counts above mirror governance/inventory.json, which is regenerated by npm run inventory and drift-checked in CI (Cycle 7 H10).

:::tip Content profiles The Standard profile (recommended) includes core, planning, implementation, review, devops, and maintenance content — everything most projects need. The Minimal profile installs only core-tagged items for a lean setup. Use hatch3r config to add or remove items after init. :::

Agents

Specialized agents that handle distinct responsibilities in your development workflow.

AgentTagsDescription
implementercore, implementationFocused implementation agent for single sub-issues, receives context from orchestrator
reviewercore, reviewSenior reviewer -- correctness, security, privacy, performance, accessibility
researchercore, planningDeep investigation sub-agent used by planning commands
fixercore, implementationTargeted fixes for reviewer findings in the review loop
test-writercore, reviewQA engineer -- unit, integration, E2E, security rules, contract tests
lint-fixercore, implementationCode quality enforcer -- ESLint, Prettier, TypeScript strict, dead code removal
learnings-loadercore, maintenanceLoad and apply project learnings from the knowledge base
context-rulescore, maintenanceDynamic context rule generation from project patterns
architectplanningArchitecture design, trade-off analysis, ADR production
docs-writermaintenanceTechnical writer -- specs, ADRs, glossary, process docs
dependency-auditormaintenance, securitySupply chain analyst -- CVE scanning, upgrade paths, lockfile integrity
security-auditorreview, securitySecurity analyst -- DB rules, cloud functions, data flows, privacy
ci-watcherdevopsCI/CD specialist -- monitors GitHub Actions, reads failure logs
devopsdevopsCI/CD and deployment operations specialist
a11y-auditorreview, a11yAccessibility specialist -- WCAG AA compliance, ARIA, color contrast
perf-profilerreview, performancePerformance engineer -- runtime profiling, bundle size, benchmarks

See the full Agents reference.

Skills

On-demand instruction bundles for specific tasks. Skills are loaded when agents need them.

SkillDescription
featureEnd-to-end feature implementation as a vertical slice
bug-fixRoot cause diagnosis, minimal fix, regression test
refactorInternal quality improvement without behavior change
releaseSemantic versioning, changelog, tagging, deploy verification
issue-workflow8-step GitHub issue workflow with sub-agent delegation
pr-creationPR conventions, branch naming, self-review checklist
incident-responseTriage, mitigation, root cause analysis, post-mortem

...and 19 more. See the full Skills reference.

Rules

Persistent coding standards and conventions always available to agents.

RuleDescription
code-standardsTypeScript strict, naming conventions, function/file limits
testingDeterministic, isolated, fast tests with clear naming
api-designVersioning, validation, error responses, auth, CORS, pagination
security-patternsInput validation, output encoding, auth, OWASP alignment
git-conventionsBranch naming, commit messages, merge strategy
deep-contextDeep context retrieval and codebase understanding for AI agents
agent-orchestrationAgent delegation patterns and multi-agent coordination protocols

...and 19 more. See the full Rules reference.

Commands

Workflows invoked inside your coding tool or via CLI.

CommandDescription
board-initCreate/connect a GitHub Projects V2 board
board-fillParse todo.md into GitHub issues with dependency analysis
board-pickupAuto-pick next issue, delegate to sub-agents, create PR
project-specGenerate project documentation from a vision
feature-planPlan a single feature with parallel researchers
test-planPlan a comprehensive test strategy with parallel researchers
releaseCut a versioned release with changelog

...and 27 more. See the full Commands reference.