Instruction file imported from thebnbrkr/agora-code (
.cursor/rules/agora.mdc). Copyright stays with the author.
agora-code — Agent Instructions
Always do this
- Before reading any file over ~200 lines, run
agora-code summarize <file>first - At session start, run
agora-code injectto load previous session context - When you discover something non-obvious, run
agora-code learn "<finding>" - Before starting any task, run
agora-code recall "<query>"to check if it's been solved before
Tool reference
| Command | When to use |
|---|---|
agora-code inject |
Session start — loads goal, hypothesis, last steps |
agora-code summarize <file> |
Before reading large files — uses DB cache when same commit |
agora-code learn "<text>" |
You find a bug, gotcha, or decision worth remembering |
agora-code recall "<query>" |
Before starting work — check if it's been solved before |
agora-code checkpoint --goal "..." |
After completing a meaningful step |
agora-code status |
Current session + DB path and row counts |
agora-code memory |
Dump all DB (sessions, learnings, snapshots, symbols); add --verbose for AST/code blocks |
agora-code list-sessions / list-learnings / list-snapshots / list-symbols / list-file-changes / list-api-calls |
View each DB table without SQL |
agora-code index <file> |
Re-index file into DB (hooks do this on edit; use if needed manually) |
agora-code scan . |
To discover API routes in the codebase |
Session lifecycle
Start → agora-code inject
Working → agora-code summarize <file> before reading large files
Finding → agora-code learn "<what you found>"
Step done → agora-code checkpoint --goal "..." --action "..."
All done → agora-code complete --summary "..."