com um clique
update-docs
// Process all markdown documentation files to generate AI summaries and embeddings (using Ollama), apply taxonomy tags, and compute KNN related links.
// Process all markdown documentation files to generate AI summaries and embeddings (using Ollama), apply taxonomy tags, and compute KNN related links.
Auto-correct relative links, asset paths, and related links in markdown documents based on folder depth.
Import YouTube video details using yt-dlp, automatically generating Astro Markdown pages with metadata, iframe embeds, and descriptions.
Ensure all markdown documents have their frontmatter titles injected as standard H1 headings at the start of the body.
Scrape, download, and localize external or relative images referenced in markdown files to Astro's src/assets folder and rewrite image links.
Full developer documentation for Astro
| name | Update Docs |
| description | Process all markdown documentation files to generate AI summaries and embeddings (using Ollama), apply taxonomy tags, and compute KNN related links. |
This skill performs a full metadata enrichment and search-engine index build for the Astro documentation collection:
gemma4:31b to read each document and generate a concise one-sentence summary.embeddinggemma to compute vector embeddings of the summaries or contents, serialized in base64 within the frontmatter.tags.txt from the project root and automatically scans document text to apply matching tags.related: key in frontmatter.gemma4:31b (for summaries) and embeddinggemma (for vector embeddings) must be pulled and active:
ollama pull gemma4:31b
ollama pull embeddinggemma
tags.txt file should exist in the workspace root containing one tag per line to configure taxonomy syncing.Use this skill:
[!TIP] The script automatically stores a
last_updated_metadatatimestamp in the frontmatter of each processed file. When executing without--force, it only updates documents whose file modification time is newer than the timestamp in the frontmatter, saving redundant API calls to Ollama.
Run the script from the project root. You can pass flags to customize execution:
Standard Incremental Run (only updates files with missing or newer/modified content since the last metadata run, syncs tags, and recalculates related links):
node .agents/skills/update-docs/scripts/update-docs.js
Force Regenerate (forces updating all summaries, embeddings, and related posts even if they already exist and have not changed):
node .agents/skills/update-docs/scripts/update-docs.js --force
Tag Sync Only (focuses only on updating and synchronizing tags based on tags.txt):
node .agents/skills/update-docs/scripts/update-docs.js --tags