Skip to content

Templates

Copy-paste artifacts for agent-driven work. Each template is opinionated and fill-in-able: placeholders in <angle brackets>, no essays. Prefer these over inventing structure mid-session — agents follow a filled template more reliably than a vague prompt.

By category

CategoryPageReach for when…
Repo docsRepo DocsStanding up or hardening agent memory files (CLAUDE.md, ARCHITECTURE.md, etc.)
Specs & planningSpecs & PlanningTurning intent into something an agent can implement without inventing requirements
ChecklistsChecklistsHuman gates: review, deploy, security, readiness
Agent operationsAgent OperationsHanding off sessions, resetting context, splitting multi-agent work
ProcessProcessADRs, releases, postmortems, architecture reviews

All templates

TemplateCategoryWhen to use
CLAUDE.mdRepo docsEvery Claude Code repo — standing orders under ~60 lines
AGENTS.mdRepo docsMulti-tool repos; source of truth when Claude Code is not the only agent
README.mdRepo docsHuman onboarding; keep agent-critical facts out of here
ARCHITECTURE.mdRepo docsSystem shape agents must not invent around
TASKS.mdRepo docsLiving work queue the agent updates in-PR
DECISIONS.mdRepo docsLightweight decision log when full ADRs are overkill
CODING_STANDARDS.mdRepo docsNon-lintable conventions agents keep violating
TESTING.mdRepo docsHow to run tests, what "done" means for verification
DEPLOYMENT.mdRepo docsDeploy path, envs, rollback — agents invent this otherwise
SECURITY.mdRepo docsThreat model pointers and never-dos for secrets/auth
PROJECT_CONTEXT.mdRepo docsProduct/domain facts that don't belong in CLAUDE.md
New project briefSpecsBefore scaffolding; forces scope and non-goals
MVP specSpecsFirst shippable slice with acceptance criteria
Feature specSpecsAny non-trivial feature into an existing codebase
Bug reportSpecsRepro + expected vs actual before asking an agent to "fix"
Refactor planSpecsBehavior-preserving change with blast radius named
Task breakdownSpecsSplitting a spec into agent-sized units
Sprint planSpecsTeam cadence with agent capacity explicit
Code review checklistChecklistsHuman review of any PR (AI or not)
AI-generated PR checklistChecklistsExtra gates for agent-authored diffs
Deployment checklistChecklistsPre-prod and go-live
Security review checklistChecklistsAuth, secrets, tenancy, injection surfaces
Testing planChecklistsWhat to verify before merge
Production readiness reviewChecklistsFirst production cut or major subsystem
Agent handoff noteAgent opsEnding a session so the next one continues cleanly
Context reset noteAgent opsMid-task thrash — compact and restart
Repository onboarding noteAgent opsFirst session in an unfamiliar repo
Multi-agent task splitAgent opsParallel agents with non-overlapping ownership
ADRProcessCross-cutting design choice that must survive sessions
Release notesProcessShipping a version humans and agents can cite
PostmortemProcessIncident or agent-caused outage write-up
Architecture review notesProcessStructured review of a proposed or existing design

32 templates. Prefer linking the filled artifact from CLAUDE.md over pasting it into every prompt.

A field manual for AI-native software engineering.