Weike Yang

    0followers

    Independent security researcher and developer. I build local-first AI security tooling: AI Content Guard is a deterministic scanner for prompt injection, hidden characters and encoded payloads. I also maintain PITAX-based MCP tooling (pitax_scan, code_audit, docs_scan). My design rule: security checks should be verifiable, low-privilege and privacy-respecting — no network, no telemetry, redacted evidence.

    1 skill
    1 use
    Joined Nov 2026

    Skills by Weike Yang (1)

    AI Content Guard — Prompt Injection & Hidden Char Scanner

    by Weike Yang

    Free

    Local, deterministic scanner that screens untrusted content — files, emails, docs, repos, AI configs — for prompt injection, invisible characters, encoded payloads and config backdoors before your agent reads it. Zero network. Redacted reports.

    1
    1