Evidence-driven deep research agent for Claude Code and Codex/ChatGPT Work, with schema v2, claim-ledger audits, safe screenshots, and strict QC.
This repository also configures its own agents. See what research-agent tells them →
Skill Claude Code
A six-stage workflow for researching a topic and producing Markdown and HTML reports with a supporting evidence table. It covers research, source checking, outlining, chapter work, editing, screenshots, and quality checks.
Skill Codex
Execute evidence-backed six-phase research in this repository and deliver Markdown, HTML, evidence matrices, data tables, screenshots, and disclosed research gaps. Use when the user requests a deep research report, a research plan, or an autonomous research run with regular, plan, or execution mode in ChatGPT Work or…
Skill Claude CodeCodex
Use when user requests systematic deep research on a topic. Six-phase pipeline (survey → evidence → outline → chapter research → assembly+editing → delivery) producing Markdown+HTML dual-format reports with evidence matrices. Enforces no-fabrication red lines, source tiering, claim-ledger audits, workflow modes, and…