Fix OpenClaw missing skill dependencies

Identifying the exact dependency gap

When a skill fails to activate due to a missing dependency, the error message in the agent logs names the specific package or command that is not found. Look for lines containing Error: or Module not found: in the activation log. The package name is usually the last quoted string on the error line.

Each OpenClaw skill declares its dependencies in a manifest.yaml file within the skill's directory. This file lists required Python packages (pip), Node.js modules (npm), and system binaries. Review the manifest for the skill you were attempting to activate and verify each listed dependency against your current installation.

Run clawmesh skills doctor to get a comprehensive report of all declared dependencies across your active skill set. This command checks for each dependency's presence and version, and reports a pass or fail status per skill. Use this output to prioritize fixes — address the skills you actively use first.

• Read the skill activation error message carefully — it names the missing package
• Check the skill manifest file (.clawmesh/skills/{skill-name}/manifest.yaml) for declared dependencies
• Run clawmesh skills doctor to scan all installed skills for missing dependencies
• Cross-reference the missing package name with the skill's documentation page

Fixing Python package dependencies

Many OpenClaw skills are written in Python and depend on packages from PyPI. If a skill fails with a ModuleNotFoundError or ImportError, the missing package can usually be installed with pip. However, avoid installing packages globally with sudo pip install, as this can create conflicts with the system Python installation.

Create a virtual environment for OpenClaw skill development and dependency management. Run python3 -m venv ~/.clawmesh/venv to create an isolated environment, then activate it with source ~/.clawmesh/venv/bin/activate before installing packages. This prevents version conflicts between different skills that may require different package versions.

On Apple Silicon Macs, some Python packages have pre-built wheels for ARM64 while others require compilation. If a pip install fails with a compilation error on Apple Silicon, check if a specific Python version is required by the package and install that version via pyenv before attempting the package installation again.

• Skills requiring Python packages use pip; install with pip install {package-name}
• Use a virtual environment to avoid conflicting package versions
• Pin installed versions in requirements.txt for reproducibility
• On Apple Silicon, ensure pip points to the ARM64 Python build

Fixing Node.js module dependencies

Some skills include custom Node.js scripts that require third-party npm packages. If a skill reports a missing Node.js module, navigate to the skill directory and run npm install to install all packages listed in its package.json. Do not run npm install globally — each skill should manage its own node_modules directory.

If you encounter peer dependency warnings during npm install, resolve them with npm install --legacy-peer-deps. Peer dependency conflicts occur when a skill's direct dependencies require different versions of a shared package. Using legacy peer resolution is safe for skill development as long as the skill's own functionality is unaffected.

After installing Node.js dependencies, restart the OpenClaw runtime to ensure the newly installed modules are loaded. Run clawmesh restart to reload the runtime without losing your workspace configuration.

• Skills using Node.js modules list them in package.json under dependencies
• Run npm install in the skill directory to install missing modules
• Use npm install --save-exact to pin versions in package-lock.json
• Avoid global npm installs for skill-specific modules

Fixing system binary dependencies

Skills that perform media processing, browser automation, or Git operations require external system binaries that are not part of the OpenClaw runtime itself. When these are missing, the error logs will show a 'command not found' message for the specific binary name.

On macOS, install missing system tools with Homebrew: brew install ffmpeg git-lfs chromium. On Linux, use apt: sudo apt install ffmpeg git. On Windows, use winget: winget install Gnu.Git. After installation, verify the binary is on your PATH by running which ffmpeg or ffmpeg --version.

The browser relay skill requires Chromium to be installed for headless browser automation. If the relay fails with a missing browser binary error, install Chromium via your system's package manager or download it directly from the Chromium project website.

• Some skills require external system binaries (e.g., ffmpeg, chromium, git)
• Check the OpenClaw logs for 'command not found' errors pointing to system tools
• Install missing tools with Homebrew (macOS), apt (Linux), or winget (Windows)
• Verify the binary is on your PATH before retrying the skill

Applying minimal remediation

The minimal remediation approach is to install exactly the missing package, validate the skill works, then stop. Avoid the temptation to run pip install everything or npm install when only one specific package is missing. Installing extra packages creates unnecessary dependencies and can introduce version conflicts.

After installing a missing dependency, run the skill activation command explicitly rather than submitting a full task. This isolates the fix validation from task execution logic. If the activation succeeds, the skill is ready for use. If it still fails, the error message will point to the next missing component.

Keep a record of every dependency you install manually. If a future OpenClaw update changes the skill manifest, your notes will help you understand what was added outside the standard installation process.

• Install only the missing package, not all dependencies at once
• Validate after each installation by running the specific skill activation
• If a broad reinstall is needed, test skills one at a time after
• Log each fix so you can revert if a later skill fails unexpectedly

Preventing future dependency issues

The Skills Hub provides managed skill installations where dependencies are pre-bundled and versioned. For production deployments where reliability matters more than customization, use managed skills from the hub rather than manually maintained skill git repositories. Managed skills are tested as a complete unit and update as a coordinated set.

If you maintain skills manually, pin the git commit or version tag for each skill in your configuration file. This prevents unexpected updates when someone pushes a change to the skill repository. Update skills deliberately, not automatically.

• Pin skill versions when you have a working configuration
• Review the skills hub changelog before updating skills
• Use the managed skills path for production deployments
• Test skill updates in a staging environment before applying to production