Blog
News, launch notes, guides, and comparisons for builders exploring reusable AI agent skills, trust signals, and ecosystem shifts.
Latest posts
Follow product updates, editorial pieces, and practical explainers that help users navigate the marketplace and adjacent agent ecosystems.
-
Keep GitHub wiki pages synced with recently merged code changes
This entry turns GitHub Next’s Agentic Wiki Writer into a documentation-maintenance workflow. The agent watches for recent merges, reads the PAGES.md template, updates only the relevant wiki pages, and opens follow-up changes when the source structure itself needs adjustment.
-
Generate Markdown tables of contents before publishing docs or README updates
Runs DocToc to insert or refresh navigable tables of contents inside Markdown files after headings already exist. Use it when an agent is preparing README or docs updates and needs reliable intra-document navigation, not when it is generating the documentation itself.
-
Normalize and filter noisy URL lists before crawling or queueing
Uses Courlan to clean, normalize, de-track, and language-filter raw URL inventories before a crawler, scraper, or analyst queue touches them. Best when an agent already has too many candidate links and needs a smaller, cleaner frontier, not a full crawling stack.
-
Reformat Markdown files into a stable house style before review or migration
Runs mdformat to rewrite Markdown into a consistent CommonMark-oriented layout or check mode in CI. Use it when an agent needs deterministic Markdown diffs before review, migration, or bulk cleanup, not when it needs to generate new content or build a docs site.
-
Split Markdown front matter and body before docs migrations, metadata audits, or static-site transforms
Use gray-matter when an agent needs to open Markdown or content files, extract structured front matter, and pass clean body text plus metadata into the next step. This skill is for parsing and rewriting front matter safely, not for running a full CMS or static site generator.
-
Generate stable URL slugs from multilingual titles before publishing pages, docs, or filenames
Use python-slugify when an agent has messy human titles and needs safe, repeatable slugs for URLs, filenames, or record IDs. This skill stays tightly focused on transliteration and normalization, not routing, redirects, or full publishing workflows.
-
Normalize vCard contact exports into structured contact records before CRM imports or dedup jobs
Use vobject when an agent receives .vcf contact exports and needs reliable parsing into names, emails, phone numbers, organizations, and addresses before import or cleanup. This skill is for vCard normalization and serialization, not contact sync, outreach, or CRM management.
-
Force-align narration and transcript text into subtitle or SMIL timing maps
Use aeneas when an agent already has audio and text, but still needs timing. The workflow aligns spoken narration against fragments of plain text or XML and emits sync maps that can be turned into subtitles, EPUB 3 media overlays, JSON timing data, or other downstream caption assets.
-
Auto-crop images around the most important subject before generating thumbnails
Use smartcrop when an agent needs to choose a sensible crop automatically instead of center-cropping every image. The workflow scores visual saliency, returns crop coordinates for a target aspect ratio, and hands those coordinates to an image pipeline that renders the final thumbnail or social card.
-
Format citations and bibliographies from DOIs, URLs, BibTeX, and CFF before publishing
Use Citation.js when an agent has raw references and needs clean citation output instead of hand-formatting sources. It can resolve supported identifiers, normalize metadata into CSL-JSON, and emit bibliography or inline citation formats that fit articles, research notes, docs, or release materials.