Instructions for ed-donner/llm_engineering: This is a Streamlit web application for clinical trial protocol summarization. Use Streamlit best practices for UI and Python for backend logic. Integrate with ClinicalTrials.gov v2 API for study search and OpenAI for summarization.
Instructions for ed-donner/tech2ai, covering from software engineer to ai data scientist, context, your role as coding agent, the topic of the live event and environment setup.
Adversarial reviewer for Personal Space. Uses the running app in unscripted, hostile ways to break it, working from the browser's text snapshot, and records every finding in ADVERSARIALREVIEW.md. Never fixes, never triages its own findings.
Backend developer for Personal Space. Implements the Node/TypeScript server, SQLite storage, API, seed data and backend unit tests from the orchestrator's task specs.
Frontend developer for Personal Space. Implements UI features and frontend unit tests from the orchestrator's task specs. Has vision — verifies its own work against screenshots before reporting done.
Delivery lead for Personal Space. Plans each phase, delegates all coding, reviews evidence, judges screenshots, triages adversary findings, and gates phases against REQUIREMENTS.md. Never writes code.
QA for Personal Space. Writes and runs the end-to-end suite, runs the full test suites, captures and inspects screenshots, and owns DEFECTS.md. Never fixes product code; only qa may close a defect.
AGENTS.md instructions for ed-donner/personal-space, covering personal space — build rules, the job, the team, repository conventions and defects.md — the defect ledger.
Claude Code instructions for ed-donner/territory, a project described as: The Territory Game for Team Competitions with Players generated by Coding Agents.