<?xml version="1.0" encoding="UTF-8"?>
<feed xmlns="http://www.w3.org/2005/Atom">
  <title>PlexusOne Blog</title>
  <link href="https://plexusone.dev/atom.xml" rel="self" type="application/atom+xml"></link>
  <link href="https://plexusone.dev/blog" rel="alternate" type="text/html"></link>
  <updated>2026-04-05T00:00:00Z</updated>
  <id>https://plexusone.dev/blog</id>
  <author>
    <name>PlexusOne Team</name>
  </author>
  <entry>
    <title>Achieving Design Consistency Across 54+ Properties with Design System Spec and Lit Web Components</title>
    <link href="https://plexusone.dev/blog/unified-design-across-platforms" rel="alternate" type="text/html"></link>
    <id>https://plexusone.dev/blog/unified-design-across-platforms</id>
    <updated>2026-04-05T00:00:00Z</updated>
    <published>2026-04-05T00:00:00Z</published>
    <summary type="text">PlexusOne spans a marketing website, 33 MkDocs documentation sites, and ~20 presentation decks. Here&#39;s how we unified design with a JSON-based spec, GitHub Pages hosting, and Lit web components replacing separate React and vanilla JS implementations.</summary>
    <author>
      <name>PlexusOne Team</name>
    </author>
    <category term="Design System"></category>
    <category term="Web Components"></category>
    <category term="Lit"></category>
    <category term="MkDocs"></category>
  </entry>
  <entry>
    <title>Multi-Agent-Spec and AssistantKit: A Growing Ecosystem for Subagents</title>
    <link href="https://plexusone.dev/blog/multi-agent-spec-assistantkit-growing-ecosystem" rel="alternate" type="text/html"></link>
    <id>https://plexusone.dev/blog/multi-agent-spec-assistantkit-growing-ecosystem</id>
    <updated>2026-02-22T00:00:00Z</updated>
    <published>2026-02-22T00:00:00Z</published>
    <summary type="text">The multi-agent-spec and AssistantKit ecosystem is maturing with multiple implementations including agent-team-release for Claude Code Agent Teams. Write agents once, deploy to Claude Code, Kiro CLI, and more.</summary>
    <author>
      <name>PlexusOne Team</name>
    </author>
    <category term="Multi-Agent"></category>
    <category term="Subagents"></category>
    <category term="Claude Code"></category>
    <category term="Agent Teams"></category>
  </entry>
  <entry>
    <title>Introducing the Releases Page: Tracking Progress Across PlexusOne</title>
    <link href="https://plexusone.dev/blog/introducing-releases-page" rel="alternate" type="text/html"></link>
    <id>https://plexusone.dev/blog/introducing-releases-page</id>
    <updated>2026-02-02T00:00:00Z</updated>
    <published>2026-02-02T00:00:00Z</published>
    <summary type="text">We built ReleaseLog to aggregate GitHub releases across all our repositories into a single, filterable view with a GitHub-style heatmap. Now live at plexusone.dev/releases.</summary>
    <author>
      <name>PlexusOne Team</name>
    </author>
    <category term="Releases"></category>
    <category term="Open Source"></category>
  </entry>
  <entry>
    <title>Why We Built Another Confluence MCP Server</title>
    <link href="https://plexusone.dev/blog/mcp-confluence-table-corruption" rel="alternate" type="text/html"></link>
    <id>https://plexusone.dev/blog/mcp-confluence-table-corruption</id>
    <updated>2024-12-29T00:00:00Z</updated>
    <published>2024-12-29T00:00:00Z</published>
    <summary type="text">Existing Confluence MCP servers corrupt tables. The root cause: LLMs generate Markdown or HTML5, but Confluence uses Storage Format XHTML. We built a server that handles both structured creation and lossless editing.</summary>
    <author>
      <name>PlexusOne Team</name>
    </author>
    <category term="MCP"></category>
    <category term="Confluence"></category>
  </entry>
  <entry>
    <title>Security-Gated MCP Servers with VaultGuard</title>
    <link href="https://plexusone.dev/blog/security-gated-mcp-servers" rel="alternate" type="text/html"></link>
    <id>https://plexusone.dev/blog/security-gated-mcp-servers</id>
    <updated>2024-12-29T00:00:00Z</updated>
    <published>2024-12-29T00:00:00Z</published>
    <summary type="text">MCP servers need API keys. Environment variables in config files are convenient but insecure. We built a pattern using VaultGuard that gates credential access on device security posture.</summary>
    <author>
      <name>PlexusOne Team</name>
    </author>
    <category term="MCP"></category>
    <category term="Security"></category>
    <category term="VaultGuard"></category>
    <category term="OmniVault"></category>
  </entry>
  <entry>
    <title>A Severity Rubric for WCAG 2.2 AA Prioritization</title>
    <link href="https://plexusone.dev/blog/wcag-accessibility-with-ai" rel="alternate" type="text/html"></link>
    <id>https://plexusone.dev/blog/wcag-accessibility-with-ai</id>
    <updated>2024-12-29T00:00:00Z</updated>
    <published>2024-12-29T00:00:00Z</published>
    <summary type="text">WCAG defines conformance levels, not severity. When you find 14 accessibility issues, which do you fix first? We built a severity rubric based on user impact and share our initial implementation pass.</summary>
    <author>
      <name>PlexusOne Team</name>
    </author>
    <category term="Accessibility"></category>
    <category term="WCAG"></category>
  </entry>
  <entry>
    <title>OpenTelemetry Semantic Conventions for Agentic AI</title>
    <link href="https://plexusone.dev/blog/otel-semantic-conventions-agentic-ai" rel="alternate" type="text/html"></link>
    <id>https://plexusone.dev/blog/otel-semantic-conventions-agentic-ai</id>
    <updated>2024-12-28T00:00:00Z</updated>
    <published>2024-12-28T00:00:00Z</published>
    <summary type="text">Multi-agent AI systems need observability beyond what OpenTelemetry GenAI provides. We built semantic conventions for workflows, tasks, handoffs, and tool calls—extending gen_ai.agent.* to give multi-agent systems first-class observability.</summary>
    <author>
      <name>PlexusOne Team</name>
    </author>
    <category term="OpenTelemetry"></category>
    <category term="Observability"></category>
    <category term="Multi-Agent"></category>
    <category term="OmniObserve"></category>
  </entry>
  <entry>
    <title>Building Production Go SDKs with Claude Opus 4.5</title>
    <link href="https://plexusone.dev/blog/ai-assisted-sdk-development" rel="alternate" type="text/html"></link>
    <id>https://plexusone.dev/blog/ai-assisted-sdk-development</id>
    <updated>2024-12-28T00:00:00Z</updated>
    <published>2024-12-28T00:00:00Z</published>
    <summary type="text">We built two complete Go SDKs in hours instead of weeks. The pattern: OpenAPI spec → ogen code generation → wrapper services. Here&#39;s what we learned about AI-assisted SDK development.</summary>
    <author>
      <name>PlexusOne Team</name>
    </author>
    <category term="Claude Opus 4.5"></category>
    <category term="SDK Development"></category>
    <category term="Go"></category>
  </entry>
  <entry>
    <title>Building a Multi-Agent Statistics Verification System</title>
    <link href="https://plexusone.dev/blog/building-stats-agent-team" rel="alternate" type="text/html"></link>
    <id>https://plexusone.dev/blog/building-stats-agent-team</id>
    <updated>2024-12-28T00:00:00Z</updated>
    <published>2024-12-28T00:00:00Z</published>
    <summary type="text">The journey of creating stats-agent-team: from hallucinated statistics to verified facts. Lessons learned building a 4-agent pipeline with OmniLLM, OmniSerp, and OmniObserve.</summary>
    <author>
      <name>PlexusOne Team</name>
    </author>
    <category term="Multi-Agent"></category>
    <category term="OmniLLM"></category>
    <category term="OmniSerp"></category>
    <category term="Case Study"></category>
  </entry>
</feed>