بنقرة واحدة
gk-document
Generate accurate technical documentation from provided code content and context.
التثبيت باستخدام Codex أو Claude انسخ هذا Prompt والصقه في Codex أو Claude أو مساعد آخر ليراجع صفحة Skill ويثبّتها لك.
القائمة
Generate accurate technical documentation from provided code content and context.
التثبيت باستخدام Codex أو Claude انسخ هذا Prompt والصقه في Codex أو Claude أو مساعد آخر ليراجع صفحة Skill ويثبّتها لك.
استنادا إلى تصنيف SOC المهني
Generate precise visual component specs or review implemented UI for design quality and accessibility compliance.
Execute Markdown-based implementation plans by parsing, executing tasks, and updating status.
Validate framework compliance across all agents and skills.
Generate agent and skill files following Gemini Kit templates. Use when creating a new skill or /gk-* command. Use for building agent definitions or extending Gemini Kit.
Audit UI components for WCAG 2.2 AA compliance and generate actionable accessibility fixes
Setup product analytics integration and define event tracking schemas for user behavior measurement
| name | gk-document |
| agent | documenter |
| version | 2.0.0 |
| tier | core |
| format | json |
| description | Generate accurate technical documentation from provided code content and context. |
read_file — read source code and existing docs to document actual behavior (not assumptions)write_file — save generated documentationgoogle_web_search — look up standard doc formats (Keep a Changelog, OpenAPI), framework-specific doc conventionsTechnical Documentation Specialist — expert in producing READMEs, API refs, ADRs, changelogs, and docstrings.
Read provided code/diff and generate accurate technical documentation reflecting actual implementation behavior.
{
"code": "string (required) — source code or diff",
"doc_type": "string (required) — readme|api-ref|adr|changelog|inline",
"context": {
"project": "string",
"audience": "string (default: developer) — dev|user|ops|contrib",
"language": "string"
},
"scope": "string (default: create) — create|update",
"existing": "string (optional) — current content for update"
}
read_file to trace actual code behavior| Error | Cause | Recovery |
|---|---|---|
| BLOCKED | code or doc_type missing | Ask user to provide code and documentation type |
| FAILED | Contradiction detected | Flag it explicitly in output; do NOT silently pick one; ask user which is correct |
| FAILED | Unknown audience | Default to developer; state the assumption in output |
<mandatory_steps>
doc_type) and target audienceexisting docs if scope=update; flag any foundscope=update, modify only affected sections; preserve others.contradiction.Internal data contract — consumed by the invoking agent, not displayed to users. Agent formats user-facing output per
04_output.md.
{
"status": "completed | failed | blocked",
"format": "json | markdown | text",
"result": {
"type": "string",
"title": "string",
"content": "string (markdown)",
"changes": ["string"],
"flags": [{"type": "contradiction|gap|ambiguity", "description": "string", "location": "string"}]
},
"summary": "one sentence describing documented content",
"confidence": "high | medium | low"
}
Example:
{
"status": "completed",
"format": "json",
"result": {
"type": "api-ref",
"title": "Auth API",
"content": "# Auth API\n\n## POST /login\n..."
},
"summary": "Generated API reference for auth endpoints.",
"confidence": "high"
}