Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Codebase Onboarding Map: Tour, Commands and a First Task

    1

    A short map of an unfamiliar repo made from file names: the commands its own manifest files name, what to open first, and one first task. It runs nothing.

    $5

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    Codebase Onboarding Map: Tour, Commands and a First Task

    Codebase Onboarding Map: Tour, Commands and a First Task

    Example session with this skill installed

    I just cloned this repo. Give me a tour: what should I read first, and how do I run the tests?

    • Read your context and instructions
    • Compiled the codebase onboarding map

    Summary

    The skill ran its own script on the example project shipped in the package (references/example) and on the failing example (references/example-failing). Output below is unedited, apart from the ... where a row is cut.

    $ python3 scripts/onboarding_map.py --project references/example --out ./map-out
    run         ok=2 fix=0 warn=0 unknown=0
    tour        ok=1 fix=0 warn=0 unknown=0
    first-task  ok=1 fix=0 warn=0 unknown=0
    map: ONBOARDING-MAP.md and onboarding-map.json in the --out folder
    MAP ok=4 fix=0 warn=0 unknown=0          (exit 0)
    

    From ONBOARDING-MAP.md:

    run: commands named in the listed files

    • ok npm test (npm-test-script; source: npm Docs): package.json has a test key under scripts. ...
    • ok npm start (npm-start-script; source: npm Docs): package.json has a start key under scripts. ...

    tour: open these first

    1. README.md 2. package.json 3. src/ 4. tests/

    first-task

    Run npm test. Then open these files and folders in order: README.md, package.json, src/.

    $ python3 scripts/onboarding_map.py --project references/example-failing --out ./map-out-2
    first-task  ok=0 fix=2 warn=0 unknown=0
    MAP ok=0 fix=4 warn=0 unknown=0          (exit 1)
    

    Next steps

    • Run the test command the map names, then open the three paths in order.
    • ok means the file, key or target line is there. It does not mean the command works or the tests pass; the skill never runs them.
    From ONBOARDING-MAP.md

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

    About this skill

    Use when you join a project and want to know where to start, when an agent should explain a repo but first needs to know which files matter, or when you want the test command before you touch anything. A standard-library script reads file and folder names and a few manifest files in the root (npm, make, Python, Go, Rust), using only their key and target names and a few header checks written out in the rules file. It writes a short map: the commands those files name (always fixed strings from closed lists, never text copied from the repo; a manifest file that cannot be read gives unknown and no command), a tour in a fixed order (README, manifest files, the largest folder, the test folder) and one first task: run the test command, then open the first three paths. When a root npm manifest that can be read names a lifecycle key from its list or has workspaces, or the root also has the build file from its list while that manifest names no install or preinstall key, the map gives warn so you read those scripts or that file first; their text is never shown. Dependencies and workspace packages are not looked at, and a root npm manifest that cannot be read (unknown) is not checked for lifecycle keys, workspaces or the build file, so no warn does not mean that installing runs no script. Every row is ok, fix, warn or unknown, and the exit code follows the rows. Do not use when you want setup scripts generated, release notes, a code review, a pull request split, an architecture analysis, a dependency check or a scan for secrets.

    Requires: Python 3.9+ (standard library only; tested with 3.9). On Python before 3.11 the Python project file is not parsed and gives unknown. You give the project folder and an output folder outside it; the scripts read names and a few manifest files in the root, follow no symbolic links inside the project folder, create the output folder (and missing parent folders) and write only ONBOARDING-MAP.md and onboarding-map.json there, start no process and make no network calls.

    Independent: Independent tool. Not affiliated with or endorsed by npm, Inc. or GitHub, nor by Anthropic, OpenAI, Anysphere (Cursor), Google, the Python Software Foundation, the Go project, the Rust project, pytest, GitLab or CircleCI. Product names are used only to describe compatibility and the files the skill looks at.

    The map is made from names, not from reading the code. ok means a file, key or target line is there; it does not mean that a command works, that the README is right or that the tests pass. Commands are suggestions from fixed lists and are never run; for Python, Go and Rust they are fixed suggestions for that kind of project. It does not explain the architecture, check dependencies, environment variables or code quality, does not look at the lifecycle scripts of dependencies or workspace packages, and does not look for secrets: a file or folder name that passes its path rule is shown as it is, so a credential in a name can appear in the map. The make rule is a line rule, not a make parser. Claims about npm come only from the npm Docs page on scripts; every other rule is marked heuristic.

    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 any payment.

    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

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed 13 days ago

    Requires

    Python 3.9+

    Permissions

    Terminal / Shell
    Read Files
    Write Files

    Reads file and folder names and a few manifest files in the project root (key and target names only). Writes only ONBOARDING-MAP.md and onboarding-map.json in the --out folder, which must be outside the project. Runs Python 3 (standard library). No network, no API keys, no environment variables, starts no process.

    Listed13 days ago

    What's inside

    Frequently Asked Questions

    No. ok means the file, key or target line the rule looks for is there. It does not mean that a command works, that the README is right or that the tests pass. The skill never runs the commands.

    The script could not decide, for example a manifest file that is empty, too large, not UTF-8, a link, or one that breaks its rule in references/rules.md. It never turns into ok.

    No. It reads names, and only key and target names from a few manifest files in the root. No text from the project's files is shown, and error messages never repeat your arguments or file text.

    No. It writes nothing in the project folder; --out is refused when it is the project folder or inside it, and an existing map file is never overwritten.

    No. A file or folder name that passes the path rule is shown as it is, so a credential in a name can appear in the map.

    No. It is a line rule: conditionals, variables in target names and pattern rules are not evaluated.

    New in Technical Documentation