LRE Deep Research
- Profile, books, investments, ideas
- Single-copy notes update flow
- Email output with stage aggregation
Self-Creation · Self-Healing · Chain Tools
AgInTi
A practical prompt-tool system for deep research, self-healing automation, and
continuous cross-repo updates. This page is fully self-contained under
website/ for stable GitHub Pages deployment.
4
LRE core stages
2
Course-agent generation phases
5
AutoAppDev repo targets
Live
Custom domain pipeline
Core Systems
AgInTi connects research, execution, and maintenance using dedicated prompt tools.
Course Agent
- Selenium bootstrap with session reuse
- Tool builder and autodev loop
- Prompt-driven recovery for stalls
AutoAppDev
- Auto-README updates
- Auto-website refresh
- Cross-repo consistency sync
LRE Pipeline
Deterministic stage chain with fallback behavior and latest-artifact promotion.
1) Profile
Identity anchors, source validation, and evidence quality checks.
2) Books
Growth-focused reading plan derived from profile strengths and gaps.
3) Investments
Thesis candidates and risk framing aligned with current objectives.
4) Ideas
Research directions, milestone draft, and self-heal checkpoints.
Course Agent Pipeline
Human-like action loop for online course completion and workflow maintenance.
A) Bootstrap
Open login session with browser automation and controlled checkpoints.
B) Build Plan
Create structured tool plan from objective and prior issue logs.
C) Autodev
Generate/fix tools inside target root and reuse runtime primitives.
D) Iterate
Use result artifacts to refine prompt logic and reduce failure loops.