grepai-troubleshooting_skill

This skill guides you through diagnosing and fixing common GrepAI issues, improving search reliability and indexing efficiently.

1

GitHub Stars

1

Bundled Files

2 months ago

Catalog Refreshed

4 months ago

First Indexed

Readme & install

Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.

Installation

Preview and clipboard use veilstrat where the catalogue uses aiagentskills.

npx veilstrat add skill yoanbernabeu/grepai-skills --skill grepai-troubleshooting

SKILL.md7.1 KB

Overview

This skill is a practical troubleshooting guide for GrepAI that helps diagnose and fix common installation, indexing, embedding, and search issues. It provides step-by-step commands, likely causes, and targeted solutions to get GrepAI back to a usable state. Use it to speed up recovery from configuration, connectivity, or performance problems.

How this skill works

The skill inspects common failure points: index presence and status, embedding provider connectivity (Ollama or OpenAI), configuration files, daemon/watch state, and symbol indexes. It recommends concrete CLI checks and corrective commands such as grepai init, grepai watch, ollama serve, model pulls, and index rebuild steps. It also includes diagnostic scripts and reset instructions for full recovery.

When to use it

When GrepAI reports "Index not found" or shows zero files/chunks
When searches return no results or irrelevant results
When embedding provider (Ollama/OpenAI) fails to connect
When indexing or search performance is slow or consumes too much memory
When MCP or trace features fail to locate symbols or integrate with assistants

Best practices

Run quick diagnostics: grepai version, grepai status, and cat .grepai/config.yaml
Keep embedder models small for limited memory and use Qdrant/Postgres for large indices
Add ignore patterns (node_modules, build, dist) to speed initial indexing
Verify Ollama or OpenAI endpoints and API keys before indexing
Restart the watch daemon and force re-index when symbols or files are out of date

Example use cases

Fix an empty search by running grepai init and grepai watch to build an index
Resolve Ollama connectivity by starting ollama serve and testing curl http://localhost:11434/api/tags
Recover from an outdated symbol index by removing .grepai/symbols.gob and restarting the watcher
Improve relevance by configuring search boosting and excluding test directories
Reduce memory use by switching to a smaller embedder model or changing backend to qdrant

FAQ

Check grepai watch --status, stop the daemon with grepai watch --stop and restart with grepai watch --background; if issues persist, remove .grepai/index.gob and rebuild.

How do I fix a 401 Unauthorized OpenAI error?

Confirm OPENAI_API_KEY is exported in your environment, verify echo $OPENAI_API_KEY, and ensure the config references the env var (embedder.api_key: ${OPENAI_API_KEY}).