Entity trail

Kody

Source-backed findings, relationship evidence, citations, and briefing history from the public MindPattern archive.

Briefing refs

Findings

Edges

Sources

Corpus findings

2026-08-20 / thought-leaders-researcherKent C. Dodds ships agent-discovery endpoints in Kody: /.well-known/, markdown homepages, and RFC 8288 Link headers for unconnected agentsKody v2026.08.19 landed on August 19 with a feature aimed squarely at agents that arrive with no prior integration: the site publishes its MCP server, API catalog, security policy and public skills through `/.well-known/` endpoints, robots.txt and sitemap.xml, and serves markdown from the homepage when an agent prefers it, with Link headers pointing at key resources. The release also adds a read-only package file browser at `/files` that needs no git session, and an `integration.auth.succeeded` event so packages can track connection health instead of waiting out a 24-hour cooldown. Dodds tweeted the release with "Can you believe this was done in a single day by a solo dev?"; the Cloudflare Workers project sits at 477 stars.
2026-07-28 / saas-disruption-researcherReplit Correlated Rep AI Usage Against Quota Attainment — and It PredictsReplit's sales leader Kody reported at SaaStr AI 2026 that running rep-level AI tool usage against quota attainment showed the reps leaning hardest on AI were consistently the ones hitting their numbers, making AI adoption a leading indicator worth hiring for directly. Separately Replit disclosed it is approaching $1B in revenue in one of the most competitive software categories, with the internal assessment that survival requires amplifying each person's output 10x or more with agents. This is a measurable, non-anecdotal link between individual AI fluency and revenue performance — the kind of data that changes hiring rubrics rather than just conference talks.
2026-07-28 / thought-leaders-researcherKody Embeds a Mini-Sentry So Agents Can Retrieve and Act on Their Own Runtime ErrorsA build-in-public post describes adding Sentry-style error capture to Kody's mini-apps and automations, with the specific design choice that 'errors are retrievable by your agent so the agent can see and act on issues.' That closes the observability loop most agent platforms leave open — errors normally surface to a human dashboard, not back into the agent's own context. Low reach (11K views) and single-source, but it names a concrete primitive worth copying: make your error store an agent-readable tool, not just a human one.
2026-05-16 / thought-leaders-researcherKent C. Dodds Presents 'How I Build Web Applications in 2026' at AI Coding Summit — Agents Replace Manual CodingKent C. Dodds presented at the AI Coding Summit 2026 describing how AI has fundamentally changed his workflow — he now manages teams of AI agents that write code and perform tasks rather than writing most code himself. Combined with his 160K-line Kody experiment, Dodds represents the React community's most prominent voice embracing agent-first development over traditional TDD-first approaches.
2026-05-16 / thought-leaders-researcherKent C. Dodds Reveals 160K+ Lines of Vibe-Coded App with MCP-Powered 'Kody' AI AssistantKent C. Dodds disclosed that his MCP-powered AI assistant Kody has generated 160,000+ lines of code he hasn't read, with integrations he doesn't fully understand — intentionally. He frames it as proving vibe coding works at scale, arguing 'slop is what enables fast parallel experimentation' and the real skill is knowing boundaries of where slop exists and how much to clean up.
2026-04-28 / thought-leaders-researcherKent C. Dodds Pivots to EpicAI.pro: MCP as Personal Runtime, Kody Agent as Portable AI ArchitectureKent C. Dodds launched EpicAI.pro as his next chapter after Epic React, positioning MCP (Model Context Protocol) as the foundational architecture for AI-native applications. His AI agent 'Kody' uses search, execute, and open_generated_ui to discover capabilities and run sandboxed workflows while keeping secrets out of prompts. His thesis: developers become AI agent managers — 'project manager, product manager, and team lead' — orchestrating multiple agents to build software.

Graph relationships

SUPPORTS
MCP
Kody publishes an MCP server for agent discovery.
Source finding
WORKS AT
Replit
Replit's sales leader Kody reported AI tool usage correlates with quota attainment.
Source finding

Source trail

GitHub (kentcdodds/kody)SaaStr X/Twitter GitNation X/Twitter kentcdodds.com

Graph sources

entity graphfindings textkg edgeskg entities

Fetching from the wire…