From prompts to workflows
+PhD-Zero decomposes research into reusable stages instead of treating it as a single response.
+diff --git a/.gitignore b/.gitignore index 14c1168..ae6804d 100644 --- a/.gitignore +++ b/.gitignore @@ -1,5 +1,9 @@ .DS_Store docs/ +!docs/ +docs/* +!docs/index.html +!docs/styles.css workspace/ WORKSPACE/ .project_local/ @@ -65,4 +69,4 @@ htmlcov/ # Pyright .pyright/ -runs/ \ No newline at end of file +runs/ diff --git a/docs/index.html b/docs/index.html new file mode 100644 index 0000000..3a73aee --- /dev/null +++ b/docs/index.html @@ -0,0 +1,164 @@ + + +
+ + +Research Infrastructure for Coding Agents
++ PhD-Zero turns AI research workflows into reusable agent skills. It gives coding agents a + disciplined operating layer for planning, evidence gathering, experiment execution, memory, + human review, and paper writing. +
+Why It Exists
+PhD-Zero decomposes research into reusable stages instead of treating it as a single response.
+Skills are built around search, verification, experiments, and explicit reporting.
+Project context and memory skills let agents accumulate reusable research state over time.
+How It Works
+.agents/skills/AGENTS.md + workspace skills.claude/skills/ mirror layerQuick Start
+git clone https://github.com/TenureAI/PhD-Zero.git
+cd PhD-Zero
+ find .agents/skills -mindepth 1 -maxdepth 1 -type d
+find .claude/skills -mindepth 1 -maxdepth 1 -type l
+ # Codex / Copilot-style agents
+cat AGENTS.md
+
+# Claude Code discovers mirrored skills
+find .claude/skills -mindepth 1 -maxdepth 1 -type l
+
+ If the skill names listed under .agents/skills and .claude/skills match,
+ the shared skill layer is wired correctly.
+
Roadmap
+Scoped literature investigation, experiment planning, execution, debugging, and report drafting.
+Persistent project context, reusable memory, better iteration loops, and stronger checkpoints.
+Independent hypothesis generation, self-directed experimentation, and long-horizon research execution.
+