Menu
CocoIndex content indexing and search — use when you need to search indexed project content or trigger re-indexing
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Based on SOC occupation classification
Long-running iterative development loops. Run arbitrarily-long tasks without diluting model attention. Triggers: ralph, ralph loop, iterative loop, long-running task, development loop.
Full compactor reference — tool parameters, anti-patterns, sandbox languages, context budget, workflows.
Diagnostics — validate config, session DB, runtimes, and troubleshoot compactor issues.
Context management — compact session, recall history, run code, search content.
Stats display — context savings, session metrics, compactions, sandbox and recall/search counters.
Helps you use the /unipi:btw side-conversation workflow effectively. Use when you want to think in parallel, ask side questions without interrupting ongoing work, or inject a side thread back into the main agent.
| name | cocoindex |
| description | CocoIndex content indexing and search — use when you need to search indexed project content or trigger re-indexing |
Use CocoIndex tools and commands when you need to:
cocoindex_search instead of content_search or ctx_search/unipi:cocoindex-update after significant code changes/unipi:cocoindex-init for new projects| Aspect | Old (FTS5) | New (CocoIndex) |
|---|---|---|
| Chunking | Heading/paragraph split | AST-aware code, recursive text |
| Search | BM25 + trigram fuzzy | Semantic vector + full-text |
| Incremental | Full re-index every time | Delta-only (changed files) |
| Scale | Small corpus | Multi-GiB, parallel |
| Store | SQLite FTS5 | LanceDB |
cocoindex_searchSearch indexed content. It uses semantic vector search when the LanceDB table has vectors, LanceDB full-text search when an inverted index exists, and a lexical fallback for older text-only indexes.
cocoindex_search({ query: "how authentication works", limit: 10 })
Returns results with title, content, source, rank, contentType, and matchLayer.
cocoindex_statusCheck indexing status — pipeline configured, last run, doc count, CLI availability.
cocoindex_status({})
| Command | Description |
|---|---|
/unipi:cocoindex-update | Run cocoindex update on current project |
/unipi:cocoindex-status | Show indexing status |
/unipi:cocoindex-init | Scaffold a default pipeline |
/unipi:cocoindex-search <query> | Search indexed codebase semantically |
/unipi:cocoindex-settings | Show configuration |
Python 3.10+ with cocoindex installed:
pip install cocoindex
pip install 'cocoindex[lancedb]'
Pipeline initialized — run /unipi:cocoindex-init once per project
Embedding config — CocoIndex reuses the memory package's embedding settings:
~/.unipi/memory/config.jsonProject files → LocalFile source → SplitRecursively → EmbedText → LanceDB
↓
Agent query ──────────────────────────────→ vector search → ranked results
.unipi/cocoindex/main.py (auto-generated, customizable).unipi/cocoindex/.lancedb/| Issue | Fix |
|---|---|
| "CLI not found" | pip install cocoindex |
| "Pipeline not initialized" | /unipi:cocoindex-init |
| "Search unavailable" | npm install @lancedb/lancedb |
| "No results" | Run /unipi:cocoindex-update first |
| Embedding errors | Check ~/.unipi/memory/config.json API key |