heilashahidi/adversarial-openemr
Multi-agent adversarial evaluation platform that continuously red-teams a live OpenEMR Clinical Co-Pilot — 5 agents, 50 seed attacks, two-tier LLM Judge (Haiku 4.5 → Sonnet 4.5), deterministic regression harness.
Multi-agent adversarial evaluation platform that continuously red-teams a live OpenEMR Clinical Co-Pilot — 5 agents, 50 seed attacks, two-tier LLM Judge (Haiku 4.5 → Sonnet 4.5), deterministic regression harness.
npx skills add heilashahidi/adversarial-openemrMulti-agent adversarial evaluation platform that continuously red-teams a live OpenEMR Clinical Co-Pilot — 5 agents, 50 seed attacks, two-tier LLM Judge (Haiku 4.5 → Sonnet 4.5), deterministic regression harness.
Shareable Claude Code infrastructure for PhD researchers — skills, agents, hooks, and rules for academic workflows
Harness engineering toolkit for AI coding agents — structured specs, task pipelines, and feedback loops for Claude Code.
A portable Codex skill for avoiding Unicode, UTF-8, and mojibake pitfalls on Windows PowerShell, especially in Chinese/CJK projects.
Fork a public GitHub repo privately and keep it auto-synced with upstream. Includes workflow + Claude Code skill.
Route AI requests across 12 providers with a token-efficient assistant that compresses prompts and manages multi-agent workflows via CLI or HTTP.
10 Claude Code skills for compound engineering workflows — session discipline, harness auditing, phase governance, and issue pipeline