agent-web-reporting

Publishes sanitized static HTML reports to a local LAN web server and returns review URLs.

Updated Apr 14, 2026
One-click install
npx skills add https://github.com/amitkarpe/agent-skills --skill agent-web-reporting-amitkarpe
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: agent-web-reporting
Source: https://github.com/amitkarpe/agent-skills/tree/main/archive/skills-20260611-html-overlap/agent-web-reporting
Command: npx skills add https://github.com/amitkarpe/agent-skills --skill agent-web-reporting-amitkarpe

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve? Turning agent run results into browser-readable reports often means ad-hoc HTML, leaked secrets, or files dumped in /tmp that nobody can open from another machine. This Skill standardizes publishing sanitized static HTML reports to a local web root (/opt/agent-web) with a stable URL contract (http://192.168.0.9/<lane>/<file>.html) so results are reviewable in desktop Chrome. ## Core Features & Use Cases - Sanitized publishing pipeline: Validates HTML sources, blocks obvious secret patterns (keys, tokens, .env content), backs up overwritten files to an archive folder, and verifies the published URL with curl. - Genesis design system: Ships self-contained report and dashboard templates (navy header, indigo accents, system fonts, no external CDNs or JavaScript) plus a DESIGN.md style guide for custom reports. - Lane organization and diagrams: Organizes reports into short lane folders (work, td, pat, etc.), regenerates lane index pages, and renders Graphviz or Mermaid diagrams locally to inline SVG for workflow and architecture explanations. - Use Case: After finishing an AMI image-builder investigation, generate a Genesis report with an inline SVG lineage diagram, publish it to the td lane, and hand Amit the URL http://192.168.0.9/td/imagebuilder-decision-20260526.html for review. ## Quick Start Ask the agent to create a sanitized HTML report of the current task results and publish it to the local web server, returning the http://192.168.0.9 review URL.

Frequently Asked Questions about agent-web-reporting

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

FAQPage Schema
How do I publish an HTML report to a local web server from an agent workflow?▼

Write the report HTML in a durable evidence folder, then run scripts/publish-html-report.sh with --source, --lane, and --slug. The script copies it to /opt/agent-web/<lane>/, validates it, and prints the LAN URL such as http://192.168.0.9/td/my-report.html.

How do I create a starter HTML report template from the command line?▼

Run publish-html-report.sh with --new, --lane, --slug, and --title, optionally adding --summary, --status, --risk, and --next flags. It renders the Genesis template into ~/.AGENTS-temp/agent-web-reporting/<lane>/<slug>.html without publishing it.

Can I include Mermaid or Graphviz diagrams in static HTML reports?▼

Yes, render diagrams locally to SVG using scripts/render-diagram.sh with --type graphviz or --type mermaid, then embed the SVG inline in the HTML. Reports must not fetch Mermaid or other libraries from a CDN at view time.

Does the publishing script check reports for secrets before publishing?▼

Yes, the helper rejects sources matching patterns like AWS secret keys, private key blocks, tokens, and password strings, and warns about missing title, h1, or timestamp markers. A separate scan script also checks the skill repo for forbidden files like .env or .pem.

What are the limitations of this static HTML reporting approach?▼

Reports are static files with no database, JavaScript interactivity, or service state, so dashboards are regenerated from published files rather than live data. It targets desktop Chrome on the local LAN and is not suited for authenticated multi-user hosting.