53j1sc

Find content in your repo that talks to AI agents behind your readers' backs — hidden HTML comments, concealed elements, invisible unicode.

Layered SD-JWT credential library for Verifiable Intent (VI) — cryptographically binds an AI agent's commercial actions to a user's delegated purchase intent.

A lightweight MCP security guard for Claude Code users and indie AI builders. Claude Codeユーザーのための、接続前MCPセキュリティ診断ツール。

A public memory-poisoning benchmark for MCP memory servers — measures ingestion-block and containment rates against a corpus of known injection techniques.

MCP server for analyst-first owned-target web security assessments of authenticated, high-friction apps

Domain services for logic that spans entities — stateless policies/calculators returning Result.

A deterministic, Lighthouse-style i18n / CJK readiness scorecard and README badge — how well a codebase handles Japanese, Chinese, Korean and Unicode text.

Client-side file threat scanner for React forms. Detects malware signatures, hidden macros, polyglot files, and suspicious payloads before upload.

Content safety engine for AI agents — detects PII, prompt injection, credentials, and malicious patterns

Defend against LLM-based web scrapers in Next.js applications

Arcis MCP server. Exposes Arcis CLI scanners (audit / scan / sca) and the prompt-injection detector as tools that Cursor and any other MCP-aware AI agent can call. One install, four tools, zero cloud round trip.

Developer-first prompt-injection scanner for CI and local use

One-click AI safety proxy with web GUI - a VPN for AI API calls

Security scan for AI agent skills, MCP server configs, rules files, and hooks — prompt injection, secret exfiltration, hidden Unicode, dangerous shell patterns, hardcoded credentials. Read-only, zero dependencies.

Pi extension for read-only Snowflake queries via the team `claude` connection, with a specialized snowflake-analyst agent.

An AI-powered border control simulation game inspired by Papers, Please. Make critical decisions as a border agent while navigating corruption, trust, and moral dilemmas across multiple story-driven scenarios.

Adversarial security testing for AI agents — OWASP ZAP for AI agents

MCP communication security filter for Claude Code — ちゃんと見てヨシッ！

Ultra safe HTTP client for critical environments in Node.js