Kilocode
Copy the reza rules file globally or per-project. Kilocode will load context on every session automatically.
reza Integration — Kilocode
Add these rules to your Kilocode configuration to enable reza context awareness.
System Instructions
This project uses reza (Universal LLM Context Database).
The database is at .reza/context.db.
ALWAYS run these commands at the start of each session:
1. reza session handoff — check for interrupted sessions from other AI tools
2. reza query — get full project overview
ALWAYS use reza for file discovery instead of ls/find/glob:
reza query --find "keyword" — find files by purpose or path
When starting work:
reza session start --llm kilocode --task "describe task"
When saving progress:
reza session save --id kilocode-XXXXXXXX --summary "..." --context "..."
When done:
reza session end --id kilocode-XXXXXXXX
Quick Reference
| Task | Command |
|---|---|
| Project overview | reza query |
| Find files | reza query --find "keyword" |
| Check handoff | reza session handoff |
| Recent changes | reza query --recent |
| Start session | reza session start --llm kilocode --task "..." |
| Save progress | reza session save --id ID --summary "..." --context "..." |
| End session | reza session end --id ID |
| Export context | reza export |
| Quick status | reza status |
Setup
# Install reza from GitHub source once:
git clone https://github.com/swebreza/reza
cd reza
pip install -e .
cd your-project
reza init
reza watch & # optional: real-time syncWhy use reza with Kilocode?
- Skip project orientation (saves 5–15 min per session)
- Pick up interrupted work from Claude, Cursor, or Aider instantly
- Find the right file in 1 query instead of 5 globs
- Share architectural decisions across all your AI tools