Code Search

Search for "$ARGUMENTS" using ChunkHound semantic search.

Configuration

1. •Read ~/.claude/skills/chunkhound-config.json — contains embedding_args (CLI flags for chunkhound) and db_dir.
2. •Read ~/.claude/skills/projects.json — maps project names to paths.

Steps

1. •

   Read the config and registry files listed above.

2. •

   Resolve the current project from your working directory by matching against the project paths in the registry. If the cwd is inside a project path, that's the project.

3. •

   Check if the DB exists. If {db_dir}/{name}.duckdb does not exist, index it first:

   bash

   chunkhound index {project_path} --db {db_dir}/{name}.duckdb {embedding_args}
   

4. •

   Run semantic search:

   bash

   chunkhound search "$ARGUMENTS" --db {db_dir}/{name}.duckdb --semantic {embedding_args}
   

5. •

   If results were found, read the relevant files to answer the user's question. Do NOT run additional grep/glob searches. Stop here.

6. •

   If no results were found, use AskUserQuestion to ask:

   • •Question: "No results found in {current_project}. Search across all projects?"
   • •Options: The first option should be "No", followed by "All projects", then one option per OTHER registered project (only those whose DB files exist), plus the implicit "Other" option.
7. •

   If the user approves, run semantic search against the selected project(s).

8. •

   Read the relevant files from the results to answer the user's question.