Skip to content
View AIagentPRO78's full-sized avatar

Achievements

Achievement: Pair ExtraordinaireAchievement: YOLOAchievement: Pull Shark

Achievements

Achievement: Pair ExtraordinaireAchievement: YOLOAchievement: Pull Shark

Block or report AIagentPRO78

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
AIagentPRO78/README.md

banner

〉 whoami

Pseudonymous security researcher and builder.

I hunt bugs across web apps, cloud, and smart contracts, and I write each one up with the receipts: transcripts, dead ends, and a repro anyone can run. Then I build the tooling I keep wishing existed.

Lately that means agentic AI: kits that put Claude, Gemini, OpenAI and open models into shared rooms, and tools that make agent spend and context visible.

〉 how I work

Flow diagram: map the surface, form a hypothesis, reproduce in a local lab, confirm live with minimal traffic, report with receipts. A failed hypothesis loops back to an earlier step and dead ends are logged.

Every hypothesis gets tested where it can't hurt anyone first. Dashed arrows are dead ends and failed hypotheses looping back; they get logged, not deleted.

〉 shipped

Project What it does Stack
cybersec-10000 40-agent, ~10,163-dimension security analysis swarm for Claude Code: static-audit and live-ops tiers, 46 scored exploit chains, 133 capability tags, and a Jev (TypeSafe System One) decisioning layer on the live path. 502 tests passed. Source private, shared on request. Python
DJ SET Neon, browser-first wall of EDM and trance legends: 88 artists, 1,532 verified tracks, 24 genres, played through official YouTube, SoundCloud, and Mixcloud embeds with a persistent deck. Live at djset.club · Source TypeScript
tokenomics See, attribute, and prove the cost and context of every Claude Code session. Reads transcripts offline: no network calls, no telemetry, zero runtime dependencies. Docs JavaScript
claude-usage-bar Native macOS menu-bar tracker for Claude Code usage: 5-hour block, today, and month, with per-model splits. Keys live in the macOS Keychain. Swift
AgentMeet agent kits
Claude Code · OpenAI · Gemini · OpenAI-compatible		 Drop-in agents for agentmeet.chat. Clone, set a token and a room, run. The compatible kit switches between Qwen, DeepSeek, Mistral, Grok, Perplexity, Groq and Ollama with one env var. TypeScript

How the AgentMeet kits fit together

Diagram: an agentmeet.chat room exchanges @mentions and replies with four agent kits: claude-code-agent (runs on the local claude CLI with your Skills and MCP servers), openai-agent, gemini-agent, and openai-compatible-agent (Qwen, DeepSeek, Mistral, Grok, Perplexity, Groq, Ollama).

What tokenomics prints (sample report)

Sample tokenomics report: $0.9030 across 42 turns, context 67% used, Opus 82% of cost and Haiku 18%, with a recommendation to route mechanical subagents off Opus and save about $0.31.

〉 currently building

  • AgentMeet — agent-to-agent chat platform · agentmeet.chat
  • cve-mcp-server — 27-tool MCP server for CVE & threat intelligence · private, in development
  • Teams Mirror — AI presence on Microsoft Teams (chat + voice) · private, in development

〉 focus

Web Security Cloud Web3 / Smart Contracts Agentic AI MCP Bug Bounty

〉 stack

Python TypeScript Rust Solidity Foundry Burp Suite Semgrep Anthropic SDK Vercel Next.js

〉 reach

Popular repositories Loading

  1. edm-hub edm-hub Public

    DJ SET — The Legends Wall: a neon, browser-first wall of EDM and trance legends. 88 artists, 1,532 verified tracks, 24 genres, played through official YouTube/SoundCloud/Mixcloud embeds. Live at dj…

    TypeScript 1

  2. cybersec-10000 cybersec-10000 Public

    40-agent, ~10,163-dimension security analysis swarm (static + live ops) with a scored 46-chain exploit planner and Jev (TypeSafe System One) decisioning. Built and verified; source available on req…

    1

  3. soundblast-privacy soundblast-privacy Public

    Privacy policy site for SoundBlast — served via GitHub Pages

    HTML

  4. motivslap-privacy motivslap-privacy Public

    Privacy policy site for MotivSlap — served via GitHub Pages

    HTML

  5. AIagentPRO78 AIagentPRO78 Public

    Profile README

  6. agentmeet-openai-agent agentmeet-openai-agent Public

    OpenAI-powered agent SDK kit for agentmeet.chat

    TypeScript