RLM Agent Skills visual guide

Which skill do I need?

Each skill is shown as a 3D pipeline: what goes in, the steps the agent takes, and what comes out. Pick a skill to watch it run, then read its pros and cons before you install it.

    • Grey block: an item
    • Green: passed or matched
    • Red: failed, or a risk
    • Amber: flagged for a look
    • Blue: the first or a recalled item
    • Coloured block: its group
    • Orange loop: a cycle
    • Teal cylinder: memory
    • Frame: one chunk

    Drag to turn the view. Click a step to jump to it. The blocks are the items. They move and change colour at each step, so you can follow one item through the skill. The dashed line marks a boundary, and a frame marks a chunk.

    How to install

    Two steps. Install the extension once, then install the skills into each project that uses them.

    Download the .vsix from this page, install it with Install from VSIX… in the Extensions view, then run RLM: Install Agent Skills and keep the defaults. The extension installs the 14 default skills in .github/skills/ and reports 23 files written.

    Step 1: install the extension

    The extension is not on the VS Code Marketplace yet. Download the package, then install it.

    1. Download skill-pack-1.0.0.vsix from the skill-pack repository.
    2. Install it from a terminal, or from VS Code.
      code --install-extension skill-pack-1.0.0.vsix
      In VS Code, open the Extensions view (⇧⌘X), choose Install from VSIX… from the … menu, select the file, and reload if asked.

    Step 2: install the skills into a project

    1. Open the project as a folder (File > Open Folder), not a single file.
    2. Run RLM: Install Agent Skills from the Command Palette (⇧⌘P on macOS, Ctrl+Shift+P on Windows).
    3. Answer the three questions.
    QuestionDefaultWhat to choose
    How should the helper scripts run?1. Python1 if Python 3 is installed. 2 if only Node.js is installed. 4 if programs may not run. Option 3 is listed, but it does not work in the standard package (see the note below).
    Where should the skills go?GitHub CopilotEvery agent you use. You can pick more than one.
    Which skills should be installed?Fourteen selectedKeep the defaults, or tick the off ones you need (the three compliance skills, release-notes, stack-auditor, trust-layer-validator, incident-postmortem).
    About option 3, Portable Python. The portable Python interpreter adds about 22 MB, so the standard .vsix leaves it out. Choosing option 3 stops the install with the message “the portable Python runtime is not bundled in this build”. Choose option 1, 2 or 4 instead. Developers can bundle it by running node scripts/fetch-portable-python.mjs in the source repository and rebuilding the package. It has been prepared for Windows, but not tested there.
    1. Wait for the message RLM: wrote N file(s). If it says some files were left unchanged, those were files you had edited.

    Where each agent reads its skills

    AgentSkills folder in your project
    GitHub Copilot (default).github/skills/
    Claude Code.claude/skills/
    OpenAI Codex.agents/skills/
    Cursor.cursor/skills/
    Gemini CLI.gemini/skills/
    Windsurf.windsurf/skills/

    Step 3: check the install

    Then ask your agent in plain words, naming the skill:

    Use the rlm-security-audit skill on src/. Report findings with file and line.

    Update, remove, or get the source

    More

    Questions

    Does it need an API key?

    No. The extension makes no model calls. The agent you use sends the code it reads to its own model provider, under your subscription.

    Does it change my code?

    It writes skill files into the agent folders in your project. It never deletes anything, and files you have edited are left alone.

    Where is memory stored?

    In .agent-memory/graph.json in the project. Add .agent-memory/ to .gitignore if the project uses git.

    Why is Portable Python listed but not working?

    It is listed in the install picker, but the standard package does not include the interpreter, because it adds about 22 MB. Choose Python 3, Node.js, or instructions only. See the note in the install section.

    Is it on the Marketplace?

    Not yet. Install it from the .vsix file.

    Before you rely on the output

    About the author

    RLM Agent Skills is built and maintained by Kamalesh (kamalsoft).