ProjectSetup

    0

    Bootstrap any new or imported repo from its README/manifests and create/maintain safe AGENTS.md guardrails and context.md project memory.

    Free

    Security scanned

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    ProjectSetup

    Example session with this skill installed

    I just cloned this repo. Can you help me get it set up and establish some rules for how we work together?

    • Read your context and instructions
    • Compiled the projectsetup

    I have inspected the repository and identified a Node.js stack using pnpm. I have created local guardrails and context files to ensure safe operation.

    Project Onboarding Summary

    ComponentStatusDetails
    Stack DetectedIdentifiedNode.js (v20+), pnpm, TypeScript
    GuardrailsCreatedAGENTS.md initialized with strict no-push/no-delete rules
    ContextInitializedcontext.md seeded with purpose and verified commands
    Setup PlanPendingpnpm install and pnpm build discovered

    Verified Commands

    
    # To install dependencies
    pnpm install
    
    # To build the project
    pnpm build
    
    # To run tests
    pnpm test
    

    Initial Agent Rules (AGENTS.md)

    
    # Agent Rules (Project)
    ## Never do
    - Push to any remote or create/merge PRs.
    - Commit to git or delete files without specific permission.
    - Share or log secret values from .env or .key files.
    
    ## Always do
    - Read README/manifests before running setup.
    - Record verified commands in context.md.
    

    Next steps

    1. Review the proposed setup commands above. Should I proceed with pnpm install?
    2. Confirm if you want AGENTS.md and context.md added to .gitignore to keep them local.
    3. Provide the project purpose if it is not clearly defined in the current README.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    ProjectSetup onboards a project the way a careful senior engineer would: read first, run only what the docs say, and leave the repo with memory and guardrails.

    What it does

    • Inspects README/CONTRIBUTING, manifests and lockfiles, CI configs, and existing AGENTS.md/context.md.
    • Detects the real stack and verifies documented install / build / dev / test commands — it never assumes Node or npm.
    • Creates or safely merges AGENTS.md: project non-doings, "ask before" actions, and "always do" rules. Existing user rules are preserved and never weakened.
    • Creates or safely refreshes context.md: purpose, what the project is supposed to do, current focus, an append-only update log, and the verified setup commands.
    • Missing or contradictory docs become questions, not invented commands.

    Safety by design

    • Never pushes, commits, opens/merges PRs, deletes files, or runs destructive DB/Docker/K8s/system commands without explicit permission for that specific action.
    • Secrets (.env*, *.pem, *.key, secret, credentials) are never read into output, stored, committed, or shared.
    • AGENTS.md and context.md are local by default — not pushed or committed unless you explicitly approve it.
    • context.md is updated only when you ask, or after you confirm a push/PR succeeded; every entry carries a date, reason, and reference.

    Works with any agent that reads the Agent Skills format: OpenCode, Claude Code, Codex, Copilot CLI, Gemini CLI, and others. MIT licensed.

    Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Agents & Orchestration.

    Cannot auto-merge PRs. Refuses to handle secret values. Requires explicit permission for remote fetches like curl or wget. Primary focus is local environment stability.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before buying a skill.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Free to use

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 7 days ago

    • Free with an Agensi account

    Software
    Git

    Permissions

    Terminal / Shell
    Network Access
    Read Files
    Write Files
    Listed7 days ago

    What's inside

    Frequently Asked Questions

    It solves 'hallucination risks' by forcing the agent to verify your actual README and lockfiles before running any commands, preventing it from guessing build steps or executing destructive scripts.

    Yes, the skill is language-agnostic and detects stacks for Node.js, Python, Go, Rust, and more by inspecting manifest files like package.json or requirements.txt.

    You receive the core project setup skill, templates for the AGENTS.md guardrail and context.md files, and the multi-phase verification logic to install into your agent.

    The skill creates a local-first AGENTS.md file that acts as a hard constraint for the model, specifically prohibiting unauthorized git pushes, file deletions, or credential sharing.

    This skill is compatible with any AI agent or IDE extension that supports custom instructions and file-system access, such as Cursor, Windsurf, or custom AutoGPT instances.

    No, the skill operates under a 'read-only' discovery phase and will only execute setup commands after you have reviewed and approved the generated plan.

    New in Agents & Orchestration

    Free