Run

run

Run a single experiment iteration. Edit the target file, evaluate, keep or discard.

Run

Run a single experiment iteration. Edit the target file, evaluate, keep or discard.

Overview

Run is a comprehensive agent skill designed for tools workflows. It provides structured guidance and automation patterns that enable AI coding agents to handle run tasks with precision and reliability across diverse project environments.

This skill encapsulates best practices gathered from production deployments and open-source communities. It covers the complete lifecycle from initial setup through advanced configuration, ensuring consistent results whether you are working on a greenfield project or integrating into an existing codebase.

Built by openclaw, this skill follows the open SKILL.md standard and is compatible with all major AI coding tools including Claude Code, Cursor, Windsurf, Codex CLI, and ChatGPT.

When to Use

Activate this skill when the user needs to:

Set up or configure run in a new or existing project

Debug issues related to run implementation

Follow best practices for tools workflows

Automate repetitive run tasks

Review and improve existing run configurations

Integrate run with CI/CD pipelines or other tools

Core Capabilities

Automated Setup & Configuration

Generates complete configuration files with sensible defaults, proper directory structure, and environment-specific overrides. Includes inline documentation explaining each configuration choice and its tradeoffs.

Intelligent Code Generation

Produces idiomatic, production-ready code following established patterns and conventions. The generated code includes proper error handling, logging, type annotations, and test scaffolding out of the box.

Debugging & Troubleshooting

Systematically diagnoses common issues by analyzing error messages, log output, and configuration state. Provides step-by-step resolution guides with explanations of root causes to prevent recurrence.

Performance Optimization

Identifies bottlenecks and applies targeted optimizations based on measured data rather than assumptions. Tracks before/after metrics and documents the rationale behind each optimization decision.

Example Prompts

Users might ask:

"Set up run for my project"

"Debug why run is failing in CI"

"Optimize the run configuration for production"

"Add run support to the existing codebase"

"Review my run setup and suggest improvements"

"Migrate from the old run approach to the latest version"

Configuration

ParameterDefaultDescription

modeautoCLI mode: interactive, batch, or daemon

output_dir./outputDirectory for generated files and artifacts

verbosefalseEnable detailed logging for debugging

stricttrueEnforce strict validation on all inputs

timeout30000Maximum execution time in milliseconds

retry_count3Number of retry attempts on transient failures

Best Practices

Start with defaults — The default configuration is optimized for the most common use cases. Override only what you need to change.

Version control everything — Keep all configuration files and generated artifacts in version control for auditability and rollback capability.

Test in isolation first — Validate changes in a sandboxed environment before applying them to shared or production systems.

Document deviations — When you override defaults or apply custom configurations, document the reason in comments or a decisions log.

Monitor after changes — After applying any configuration change, monitor system behavior for at least one full cycle to catch unexpected regressions.

Keep dependencies updated — Regularly update dependencies and check for deprecation notices to avoid security vulnerabilities and compatibility issues.

Common Patterns

Quick setup

npx skills add openclaw/run

Verify installation

skills verify run

Run with custom config

skills run run --mode=auto --verbose

Troubleshooting

IssueCauseSolution

Skill not foundPackage not installed or path incorrectRun npx skills add openclaw/run to reinstall

Configuration errorInvalid parameter values or missing required fieldsRun skills validate run to check config

Timeout exceededOperation taking longer than configured limitIncrease timeout parameter or optimize the operation

Permission deniedInsufficient access to target files or directoriesCheck file permissions and ensure write access to output directory

Integration Guide

Follow these steps to integrate Run into your workflow:

Install the skill using your preferred package manager (npx, bunx, or pnpm)

Initialize configuration by running the setup wizard or copying the default config

Customize settings based on your project requirements and team conventions

Add to CI/CD by including the skill invocation in your pipeline configuration

Set up monitoring to track skill execution results and catch failures early

Output Format

This skill produces structured output in the following format:

{ "status": "success", "skill": "run", "version": "1.0.0", "results": { "files_generated": 3, "warnings": [], "metrics": { "duration_ms": 1250, "memory_mb": 45.2 } } }

Advanced Usage

For power users and complex scenarios:

Chaining skills — Combine this skill with related skills for end-to-end workflows using the skills chain command

Custom templates — Override default templates by placing custom files in the .skills/templates/ directory

Environment variables — Configure behavior via environment variables prefixed with SKILL_ for container-friendly deployments

Hooks — Register pre/post execution hooks to run custom logic before or after the skill executes

Dry run mode — Use --dry-run flag to preview changes without applying them

Related Skills

Skills that work well alongside Run:

Browse more skills in the Tools category

Check the full category listing for complementary tools