ck:ck-help

Skill discovery uses fuzzy matching and synonym expansion to scan local directories and map natural language queries to categorized skills. It returns matching skills or categories, enabling developers to find relevant workflows quickly without memorizing exact command names.

Yes, non-English arguments are automatically translated to English before script execution. The intent detection mechanism classifies the translated input and emits structured output markers to guide the response with relevant command details and task recommendations.

Provide a short query describing your task, such as "fix CI build," to receive recommended skills, workflow steps, and command usage. The system parses SKILL.md frontmatter and detects intent to classify whether you need an overview, category guide, or specific command details.

Intent detection classifies input into five types: overview, category, command, task, or search. Based on the classification, it emits presentation markers that guide the response format and invokes local Python scripts to produce structured outputs for the matched queries.

No external dependencies are required. The skill bundles its own Python scripts and operates by scanning the local .opencode/skills directory to index available commands and workflows for interactive discovery and execution.

Run the bundled script with a targeted query to retrieve command-level documentation. The system parses SKILL.md frontmatter from indexed skills and returns structured outputs including command details, category workflows, and search results.