save-artifact

Saves HTML artifacts to a local inbox and publishes them to a deployed gallery.

Updated Jun 8, 2026
One-click install
npx skills add https://github.com/shawn-sandy/agentics-kit --skill save-artifact-shawn-sandy
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: save-artifact
Source: https://github.com/shawn-sandy/agentics-kit/tree/main/kit/plugins/social-media-tools/skills/save-artifact
Command: npx skills add https://github.com/shawn-sandy/agentics-kit --skill save-artifact-shawn-sandy

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Claude-generated HTML artifacts live only inside the chat session, making them hard to keep, share, or publish. This Skill copies an artifact page into a dated local inbox and publishes it to a deployed artifacts gallery so it can be committed and served via GitHub Pages. ## Core Features & Use Cases - Multi-source resolution: Accepts a claude.ai artifact URL (fetched via WebFetch), an explicit local .html path, or an artifact just generated in the current conversation. - Security scrub gate: Runs a blocking secrets scan on the source file before anything is copied or published, preventing credentials from becoming publicly served content. - Dated, collision-safe publishing: Saves under a <name>-YYYY-MM-DD.html filename in .claude/artifacts/, then runs build-artifacts-index.sh to publish into docs/artifacts/ and rebuild the gallery index. - Use Case: After building an interactive HTML report in a Claude session, ask to save it — the Skill scrubs it, files it under a dated name, and publishes it to the gallery for sharing. ## Quick Start Ask the assistant to save the artifact you just built, or provide a claude.ai artifact URL or a path to a local .html file to save and publish it.

Frequently Asked Questions about save-artifact

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I save a Claude artifact to my local project?▼

Provide the artifact's claude.ai URL, a local .html file path, or reference an artifact just generated in the conversation. The Skill copies it into .claude/artifacts/ under a dated filename and publishes it to docs/artifacts/.

How do I fetch a claude.ai artifact URL from the command line?▼

Use WebFetch rather than curl or wget. Shell fetches receive the SPA shell or a Cloudflare 403 because only WebFetch carries the claude.ai login session needed to retrieve the raw artifact HTML.

Does saving an artifact check for secrets before publishing?▼

Yes, a blocking security-scrub gate scans the source file for secrets before any copy or publish step. If the gate returns BLOCKED or CANCELLED, the workflow stops and nothing is saved or published.

What happens if plan-agent is not installed when saving an artifact?▼

The artifact is still saved to the gitignored .claude/artifacts/ inbox, but it is not published. You must run build-artifacts-index.sh later to copy it into docs/artifacts/ and rebuild the gallery index.

Why does fetching an artifact URL sometimes return plain text instead of HTML?▼

Deleted or unshared artifacts return a plain sentence, and malformed URLs trigger a prose summary fallback. The Skill treats the absence of a doctype declaration as the failure signal and stops without saving a partial page.