
Grepai Troubleshooting
Diagnose GrepAI index, embedding, and search failures with copy-paste commands and fixes when semantic code search misbehaves locally.
Overview
grepai-troubleshooting is an agent skill for the Operate phase that diagnoses and fixes common GrepAI index, embedding, and configuration errors using scripted diagnostics.
Install
npx skills add https://github.com/yoanbernabeu/grepai-skills --skill grepai-troubleshootingWhat is this skill?
- Quick diagnostic block: grepai version, status, Ollama curl check, and .grepai/config.yaml review
- Documented fixes for “Index not found” via init and grepai watch
- Embedding provider troubleshooting for Ollama endpoint, firewall, and curl /api/tags
- “Model not found” remediation path for missing embed models
- Structured issue sections with symptom, cause, and solution per failure mode
- 4-command quick diagnostics block (version, status, Ollama tags curl, config cat)
Adoption & trust: 491 installs on skills.sh; 17 GitHub stars; 2/3 security scanners passed (skills.sh audits); trending (+100% hot-view momentum).
What problem does it solve?
GrepAI search fails or returns junk and you do not know whether the index, Ollama, embed model, or config.yaml is the root cause.
Who is it for?
Solo builders running GrepAI locally with Ollama or a custom embedder who hit index-not-found or connection errors mid-session.
Skip if: Greenfield adoption of GrepAI with no install yet—use setup/onboarding skills first—or production incidents unrelated to the GrepAI CLI and local index.
When should I use this skill?
GrepAI not working as expected, search returning poor results, index not updating, or connection/configuration errors.
What do I get? / Deliverables
You get a verified index (watch/init), a reachable embedding provider, the correct model loaded, and search working again without guessing at CLI steps.
- Working index after watch/init
- Validated embedder connectivity
- Corrected configuration notes for the agent session
Recommended Skills
Journey fit
Operate is the right shelf because the skill exists for when GrepAI is already in use and something breaks in search, indexing, or provider connectivity. Errors subphase matches explicit triggers: poor results, index not updating, and connection or configuration errors.
How it compares
Use as a procedural runbook for GrepAI failures instead of generic “debug my search” chat without version, status, and config checks.
Common Questions / FAQ
Who is grepai-troubleshooting for?
Developers and agent users who already use GrepAI in a repo and need step-by-step fixes when indexing or embedding providers break.
When should I use grepai-troubleshooting?
Use it in Operate when GrepAI is not working as expected, search quality drops, the index will not update, or you see connection and configuration errors during daily coding.
Is grepai-troubleshooting safe to install?
The skill only documents local CLI and curl diagnostics; review the Security Audits panel on this Prism page before running commands an agent suggests in sensitive environments.
SKILL.md
READMESKILL.md - Grepai Troubleshooting
# GrepAI Troubleshooting This skill provides solutions for common GrepAI issues and diagnostic procedures. ## When to Use This Skill - GrepAI not working as expected - Search returning poor results - Index not updating - Connection or configuration errors ## Quick Diagnostics Run these commands to understand your setup: ```bash # Check GrepAI version grepai version # Check project status grepai status # Check Ollama (if using) curl http://localhost:11434/api/tags # Check config cat .grepai/config.yaml ``` ## Common Issues --- ### Issue: "Index not found" **Symptom:** ``` Error: Index not found. Run 'grepai watch' first. ``` **Cause:** No index has been created for this project. **Solution:** ```bash # Initialize if needed grepai init # Create the index grepai watch ``` --- ### Issue: "Cannot connect to embedding provider" **Symptom:** ``` Error: Cannot connect to Ollama at http://localhost:11434 ``` **Causes:** 1. Ollama not running 2. Wrong endpoint configured 3. Firewall blocking connection **Solutions:** 1. Start Ollama: ```bash ollama serve ``` 2. Check endpoint in config: ```yaml embedder: endpoint: http://localhost:11434 # Verify this ``` 3. Test connection: ```bash curl http://localhost:11434/api/tags ``` --- ### Issue: "Model not found" **Symptom:** ``` Error: Model 'nomic-embed-text' not found ``` **Cause:** The embedding model hasn't been downloaded. **Solution:** ```bash # Download the model ollama pull nomic-embed-text # Verify ollama list ``` --- ### Issue: Search returns no results **Symptom:** Searches return empty or very few results. **Causes:** 1. Index is empty 2. Files are being ignored 3. Query too specific **Solutions:** 1. Check index status: ```bash grepai status # Should show files > 0 and chunks > 0 ``` 2. Verify files are being indexed: ```bash # Check ignore patterns in config cat .grepai/config.yaml | grep -A 20 "ignore:" ``` 3. Try broader query: ```bash grepai search "function" # Very broad test ``` --- ### Issue: Search returns irrelevant results **Symptom:** Results don't match what you're looking for. **Causes:** 1. Query too vague 2. Boosting not configured 3. Wrong content indexed **Solutions:** 1. Improve query (see `grepai-search-tips` skill): ```bash # Bad grepai search "auth" # Good grepai search "user authentication middleware" ``` 2. Configure boosting to penalize tests: ```yaml search: boost: enabled: true penalties: - pattern: /tests/ factor: 0.5 ``` 3. Check what's indexed: ```bash grepai status ``` --- ### Issue: Index is outdated **Symptom:** Recent file changes aren't appearing in search results. **Causes:** 1. Watch daemon not running 2. Debounce delay 3. File not in indexed extensions **Solutions:** 1. Check daemon status: ```bash grepai watch --status ``` 2. Restart daemon: ```bash grepai watch --stop grepai watch --background ``` 3. Force re-index: ```bash rm .grepai/index.gob grepai watch ``` --- ### Issue: "Config not found" **Symptom:** ``` Error: Config file not found at .grepai/config.yaml ``` **Cause:** GrepAI not initialized in this directory. **Solution:** ```bash grepai init ``` --- ### Issue: Slow indexing **Symptom:** Initial indexing takes very long. **Causes:** 1. Large codebase 2. Slow embedding provider 3. Not enough ignore patterns **Solutions:** 1. Add ignore patterns: ```yaml ignore: - node_modules - vendor - dist - build - "*.min.js" ``` 2. Use faster model: ```yaml embedder: model: nomic-embed-text # Smaller, faster ``` 3. Use OpenAI for speed (if privacy allows): ```yaml embedder: provider: openai model: text-embedding-3-small parallelism: 8 ``` --- ### Issue: Slow searches **Symptom:** Search queries take several seconds. **Causes:** 1. Very large index 2. GOB storage on large codebase 3. Embedding provid