Extend OpenClaw capabilities with skills

Skills are modular extensions that enhance the base OpenClaw agent with specialized task capabilities. Each skill packages the prompts, tool definitions, and runtime dependencies required for a specific capability category. The base agent provides core reasoning and task decomposition, while skills add domain-specific actions.

Available skills cover common automation scenarios including web search and content extraction, file format conversion and processing, code execution and repository operations, API integrations with popular services, and custom workflows for specific business processes. The skills catalog provides detailed documentation for each available module.

Skills load during agent startup and remain available throughout the agent session. Each skill operates within its own execution context, with isolation between skill dependencies to prevent version conflicts. When a skill requires a specific dependency version, that dependency loads in an isolated environment without affecting other skills.

The skills catalog lists all available modules with their capability descriptions, dependency requirements, and configuration options. Review the catalog to identify skills relevant to your use case before initiating installation. Installing skills you will not use adds unnecessary dependency overhead and increases startup time.

Selection criteria for skills should include task relevance (does the skill directly support workflows you need), dependency footprint (what additional packages does the skill require), and maintenance status (is the skill actively maintained with recent updates). Avoid installing skills that seem potentially useful without a specific planned use.

Some skills depend on other skills for their functionality. These dependency chains are documented in the skills catalog. When installing a skill with dependencies, the installation process automatically includes required dependent skills, but verify the complete dependency tree to understand the full scope of what is being added.

Install skills one at a time rather than batching multiple installations. Incremental installation isolates failures to specific modules and prevents cascading issues where one failed installation blocks subsequent operations. After each skill installs successfully, run the skill health verification before proceeding to the next installation.

Execute the skill install command with the specific skill identifier from the catalog. The installation process downloads required packages, verifies dependency compatibility, and activates the skill in your agent configuration. Monitor the installation output for any warnings about dependency version conflicts or missing optional components.

After installation completes, run the skill verification command to confirm the module loads correctly and its dependencies are available. Verification failure indicates either missing dependencies that require manual resolution, or incompatible version conflicts between the new skill and existing installed modules.

Skills declare their runtime dependencies in a manifest file that the installation process uses for dependency resolution. Conflicts occur when two installed skills require different versions of the same dependency package. The installation process attempts to resolve conflicts automatically by selecting compatible versions, but may fail when conflicts are irreconcilable.

For production deployments, pin skill versions after installation to prevent unexpected updates from introducing regressions. Version pinning locks the installed version and prevents automatic updates that could change behavior. Review update notifications and test updates in non-production environments before applying them to production.

The skill update command checks for available updates to installed skills and applies them if you confirm. Always review the changelog for skill updates before applying, and maintain a rollback procedure in case updates introduce regressions. Rollback involves reinstalling the previous pinned version using the version specification syntax.

Each skill supports configuration parameters that customize its behavior for your specific environment and use case. Configuration options are documented in the skills catalog with descriptions of available settings and valid value ranges. Access skill configuration through the management interface or by editing the skills configuration section in your agent settings.

Common configuration parameters include API credentials for services the skill interacts with, default behavior flags that control how the skill processes requests, and output formatting options that determine how results are presented. Some skills also support custom templates that define task-specific prompts using the skill's capabilities.

When customizing skill behavior, make incremental changes and test each modification before proceeding to the next. Document your configuration choices and the rationale for specific settings to maintain clarity when troubleshooting or migrating configurations across environments.

Common skill issues include missing dependencies, version conflicts, and initialization failures during agent startup. Missing dependencies typically occur when the skill manifest does not include all required packages or when package registry access is blocked by network policies. Verify package registry connectivity and retry installation with verbose logging for detailed error output.

Version conflicts between skills require either finding alternative skill versions that share compatible dependencies, or removing conflicting skills to achieve a compatible combination. The skills catalog indicates which skills have known conflicts with other modules.

Runtime initialization failures indicate the skill code encountered an error during loading, often caused by incompatible configuration settings or missing external resources. Check the agent log files for initialization error messages that identify the specific failure cause and remediation guidance.