sickn33/agentic-awesome-skills v19.12.0

on GitHub

[19.12.0] - 2026-10-20 - "Nine Reviewed Additions And A Pointer Repin"

Installable skill library update for Claude Code, Cursor, Codex CLI, Gemini CLI, Antigravity, and related AI coding assistants. Adds 9 catalog skills and ships 2,724 skills.

Start here

• Install: npx agentic-awesome-skills@19.12.0
• Choose your tool
• Best skills by tool
• Bundles
• Workflows

Added

• apify-linkedin-jobs-scraper - search and export LinkedIn job postings with salary, seniority, recruiter and company fields through the author's Apify Actor, with a cost estimate and an explicit yes before a large paid run. risk: critical because runs spend the user's Apify credit.
• apify-tiktok-scraper - collect TikTok videos, profiles, hashtags, comments and followers with engagement statistics through the author's Apify Actor, keeping paid AI add-ons off unless the user asks. risk: critical for the same paid-run reason.
• apify-facebook-ads-library-scraper - research public Meta Ad Library ads by advertiser or keyword, and state plainly that spend and impressions exist only for political and social-issue ads and reach only for EU/UK ads. risk: critical for the same paid-run reason.
• deed-page - publish an HTML file or static folder to a permanent public https://<slug>.deed.page/ URL with one curl request, confirming first that no secret or personal data goes public and keeping the returned token out of commits. risk: critical because publishing is public and the token can delete the site.
• design-token-contrast - measure WCAG 2.2 contrast between a project's color tokens in the light theme and dark mode, and propose the nearest passing color that keeps the hue, with the checker shipped as reviewable source.
• taasky - negotiate meeting times, prices and dates with another person or their agent through the hosted Taasky MCP server, where the human confirms every accept, reject or withdraw in Telegram. risk: critical because it acts toward another party.
• uiverify-triage-visual-changes - triage a UI Verify visual-regression build over the hosted MCP server: read the judge and the reviewers' comments, compare before and after, and only then accept baselines. risk: critical because accepting rewrites the project's baselines.
• unpaid-intern - keep a plain-markdown second brain for work and turn transcripts, docs and tickets into briefings, meeting prep, debriefs and status tables, with drafts only and grading of guesses. risk: critical because it stores work records.
• goal-instruction-writer - preview, draft, refine or audit durable agent Goals with acceptance evidence, drift controls, proportional verification and bounded waiting.

Changed

• 17 official Beatra pointers repinned - ai-headshot-studio, ai-image-generation-studio, ai-logo-maker, ai-multilingual-dubbing, ai-photo-restyler, ai-podcast-voiceover, beatra-ai-video-studio, beatra, ecommerce-listing-image-set, music-generation-studio, poster-design-studio, product-photo-studio, suno-lyrics-to-song, talking-avatar-video, viral-video-teardown-remake, voice-cloning-studio and voiceover-narration-studio now pin a source commit and a Git tree ID and fetch through git clone --filter=blob:none with a TREE-OK check, replacing the CDN zip download and the self-updater it carried.
• datacircle - rewritten from a paid-lookup REST/MCP guide into a description of the plan-pricing dataset the site publishes, so the skill downloads datacircle.json/.csv and queries it instead of describing per-call pricing.
• krexa and arcmira-find-quotes - upstream pin refreshed to 1.3.3 and the MCP connection requirement declared in the skill setup metadata.
• llm-cost-optimization, llmops-platform-engineering and multi-tenant-llm-hosting - name Bifrost beside LiteLLM in the gateway, cost and multi-tenant guidance.

Merge batch

• Ten open PRs were triaged, all of them touching catalog skills. All ten merged: the nine additions above plus the Beatra repin, each refreshed onto current main after review.
• design-token-contrast shipped its checker as skills/design-token-contrast/scripts/contraste.mjs, which the fork intake classifies as new_unapproved_path/new_unknown_extension. The maintainer repair moved it to references/contraste-checker.md as reviewable source, matching how story-bible-continuity ships its checker; behaviour is unchanged and the extracted block was run against a fixture.
• Where the automated review did not produce a passing result, approval used the exact-head maintainer attestation path, bound to the full 40-character head SHA. manual-review-required is recorded as such and is not a Tessl pass.

Review scope

Every addition was reviewed in full at its exact submitted bytes, including bundled references. The extracted design-token-contrast checker was run on a two-token fixture (reported 3.54:1 as failing, exit 1). All 17 Beatra source-commit/tree-ID pairs were verified live against their upstream repositories before merge. The paid Apify skills were reviewed statically; none was run against a live Apify account during review. The broader catalog audit remains unfinished.

Thanks

Thanks to @harshmaur for the three Apify Actor skills, @marina-aronchik for taasky, @wu736139669 for goal-instruction-writer, @f-o-x11 for deed-page, @igrlk for uiverify-triage-visual-changes, @kevinmmiddleton for unpaid-intern, @Clagoss for design-token-contrast, @beatra-ai for the pointer repin, @waynehamadi for the datacircle update and @Swpn0neel for the gateway wording, and to the upstream authors whose licensed material is included.