Evidence-backed UI delivery for coding agents working on existing Web products: declared criteria, criterion-bound artifacts, findings linked to declarations, and repair guidance. Evaluator review does not replace human semantic approval.

Automatic strong/cheap model-tier routing for DeepSeek Harness: intent-gated tier landing on the agent/request waterfall, plan-mode handoff (strong plans, cheap implements), deterministic high-risk guards on tools/pre-execute, failure escalation with TTL

Claude Code delegates coding to isolated CLI agents it doesn't trust (Codex, OpenCode, Pi, Pythinker), freezes what they produce, verifies it independently, and merges only what a human approves.

Permission-first rules, agents, skills and workflows for AI coding agents. One source → native CLAUDE.md, AGENTS.md and GEMINI.md for Claude Code, OpenAI Codex and Google Antigravity. Claude Code plugin + one-line install.

Blitz Strike — a universal MCP security-audit toolbelt. Reconnaissance at speed. Analysis in depth. Validation before report.

Live LinkedIn people, company, post and job data, ICP search and job-change signals for AI agents.

A /human-review skill for Claude Code: assembles one page that lets a human review a change set fast — diagram deltas, Code City, a feature video, complexity delta, and deep-linked snippets.

Requires dsh >= 0.1.7 (session format 4); on dsh 0.1.5 use dsh-repeat-tool-breaker@0.7. Escalating repeat detection for dsh: a synchronous ctx.tools.guard gate that counts semantic fingerprints (exact/cmd/net/host/sink) over a per-agent sliding window, wa

Plan startup directory submissions and verify the links went live, with your own AI agent.

Self-hosted URL- and file-to-Markdown service for humans and AI agents - web pages, documents, images, audio, YouTube. PWA + REST + MCP + Claude Code skill, Reddit-aware, refreshable share links.

Your agent grades every slip A+ to F-, links your books and tracks your record. No wager placement.

The knowledge compiler. Raw sources in, interlinked wiki out. Inspired by Karpathy's LLM Wiki pattern.

SRE thinking applied to Claude Code, based on Boris Cherny's Q&A. It enforces a strict 4-stage pipeline with gates between each step. You literally cannot skip verification. You cannot ship without approval.

Cognitive browser automation that thinks like your users—and helps AI agents navigate too. Simulate real user cognition with abandonment detection, constitutional safety, chaos engineering, and UX friction discovery. Sites that pass CBrowser's cognitive t

Original art from galleries and artists: search, details, delivered prices and checkout links.

Lets AI agents turn a UPI QR code plus cart details into a single review-and-pay page link that opens GPay, PhonePe or Paytm pre-filled, so a human pays in one tap. Agents can also poll live payment status and set or confirm it, closing the checkout loop with signed webhooks.

Enables Cursor and Codex agents to audit Agent Skill structure (frontmatter, broken links, missing files), analyze dependencies for undeclared, unused or vulnerable packages via OSV, and run 28 local bug and code-quality rules with SonarQube-style A–E ratings. It can also optionally pull issues, metrics and quality-gate status from SonarQube or SonarCloud, or scan and upload results.

TypeSafe Jev for DeepSeek Harness, the Model Context Protocol, and plain Node: typed judgments instead of prose, offline by default.

Harness the power of local LLMs with this TUI MCP Client for Ollama. Featuring all core MCP primitives (tools, prompts, resources), agent mode, multi-server, model switching, streaming responses, human-in-the-loop, thinking mode, model params config, system prompts, and saved preferences.

Enables AI agents to discover and work with a user's permitted Excalidraw boards through a single MCP endpoint, reading and adding elements, images, and snapshots while respecting owner, view, edit, and guest-link access controls.

Enables a coding agent to read a freeform Excalidraw sketch of a UI and build it in the user's project with their chosen component library, then draw its changes, notes, and links back onto the canvas. Lets users target parts of a drawing, adjust the project setup, and wait for updates over streamable HTTP, all locally.

Enables LLM-powered developer environments to discover and evaluate potentially beginner-friendly GitHub issues using explainable heuristics, clear rationale, and direct evidence links over stdio.

Enables AI agents to run free, no-signup Ethereum on-chain lookups over a remote HTTP endpoint: wallet ETH balance, contract/EOA detection, outbound transaction counts and simple risk heuristics, plus NFT metadata (name, description, image URLs, IPFS gateway links and raw tokenURI/uri JSON) for ERC-721 and ERC-1155 tokens. Reads only, with cached results and no key, account, or wallet connection required.

Enables coding agents to surgically harvest licensed upstream code, scaffold adapters behind canonical contracts, audit seams, verify live behavior, and promote or mount proven components while preserving the host architecture.