Agents Pack — shared skills and portable project memory for coding agentsSkip to content

Open source · Local-first CLITeach every coding agent how you work.<br>Agents Pack maintains one set of instructions, skills, and subagents, then renders them in each agent's native format. It also gives Claude Code, Codex, and Cursor portable, repository-owned project memory — without an account, a server, or anything overwritten behind your back.<br>Install — macOS and Linux$curl -fsSL https://farfarawaylabs.github.io/agentspackai/install.sh | sh

Get startedRead the docs

Works withClaude Code · Codex · Cursor<br>In the core pack27 skills · 6 subagents · 1 instruction set<br>Current pack0.27.0

You maintainOne canonical source<br>Core instructions<br>Skills<br>Subagents<br>Official pack content, plus anything you author yourself.

Agents Pack rendersClaude Code<br>.claude/rules/agents-pack/<br>.claude/skills/<br>.claude/agents/

Codex<br>AGENTS.md (managed block)<br>.agents/skills/<br>.codex/agents/

Cursor<br>.cursor/rules/agents-pack/<br>.cursor/skills/<br>.cursor/agents/

Repository-scope paths. Agents Pack records a hash for every file it writes, so a hand edit is reported as drift instead of being silently replaced.

Claude Code<br>Codex<br>Cursor

Inside your repository.agents-pack/memory/<br>Plain Markdown<br>MEMORY.mdCompact index of active shared memory

shared/Git-trackable<br>Project facts, decisions, workflows, and pitfalls for the team.

local/Git-ignored<br>Personal preferences, machine paths, and checkout-specific details.

No memory CLI · No database · No background process

01Cross-agent recall<br>Core instructions prompt each agent to recall relevant memory during orientation and save verified, durable learning at natural checkpoints.

02Repository-owned<br>Memory stays in ordinary Markdown that you can search, review, edit, and merge with the filesystem and Git tools you already use.

03Shared or local<br>Team knowledge is Git-trackable by default. User preferences, machine paths, and checkout details stay in an ignored local scope.

See how portable memory works

01<br>Install the CLI<br>One standalone executable with its checksum verified, placed in ~/.local/bin without sudo. No Bun, Node.js, or npm required.<br>$curl -fsSL https://farfarawaylabs.github.io/agentspackai/install.sh | sh

02<br>Initialize your project<br>Choose repository or global scope, the agents you use, and the components you want. Agents Pack shows the exact plan before it writes anything.<br>$agents-pack init

03<br>Just work<br>Each agent discovers the rendered files through its own native mechanism. Nothing to wire up, no extension to install, no process to keep running.

WorkflowsDebug systematically<br>Trace reproducible failures to evidence-backed root causes and verify fixes.<br>ap-debugWorkflowsReview a plan<br>Challenge a plan for mistakes, omissions, assumptions, risks, and simpler alternatives.<br>ap-review-planSecurityRun a security audit<br>Inspect concrete attack surfaces, validate findings, and report actionable security risks.<br>ap-security-auditBackendDevelop APIs<br>Build thin, secure, documented API layers over reusable business and data logic.<br>ap-develop-apisFrontendDesign distinctive frontends<br>Create intentional, production-grade interfaces and maintain the project design system.<br>ap-frontend-designBackendWrite database queries<br>Write correct, secure, performant queries with sound transaction and indexing choices.<br>ap-write-database-queriesWorkflowsStart a development session<br>Orient a new session and isolate mutable local work in a dedicated Git worktree.<br>ap-start-dev-sessionPlanningCreate a product requirements document<br>Interview the user, shape the product, and write a clear evidence-aware PRD.<br>ap-create-prdManagementManage Agents Pack<br>Manage official and user-owned cross-agent components through the safe lifecycle CLI.<br>ap-manage-agents-pack<br>SubagentsSpecialized agents rendered into each provider's native definition format, for delegated review, implementation, and research work.

Code reviewerap-code-reviewer<br>Backend Python developerap-backend-python-developer<br>Backend TypeScript developerap-backend-typescript-developer<br>UX enhancerap-ux-enhancer<br>Trend researcherap-trend-researcher<br>UX researcherap-ux-researcher

Browse all 27 skills·See the core instruction set

Preview is honest<br>--dry-run runs the real planner and writes nothing — no provider files, no installation state, no cache.

Unexpected edits stop the operation<br>Agents Pack records a SHA-256 hash for every file it writes. If one changed underneath it, the command stops and tells you which file.

Shared files stay shared<br>Codex instructions live in a marked block inside AGENTS.md. Everything outside that block is yours and is never touched.

Writes are transactional<br>A failed operation restores the previous filesystem state, and rerunning an interrupted command recovers before it replans.

agents-pack statusRead-only. Reports the health of every managed output.

cleanMatches what Agents Pack last wrote<br>missingExpected managed content is gone<br>modifiedContent differs from...