How do I index the current project?

Use /eidetic:index with no path argument to index the current project.

Can I index a specific path?

Yes, use '/eidetic:index /path/to/project' to index a specific path, or provide the path at the Detect Path step.

How do I verify indexing status?

After indexing, call get_indexing_status(path=...) to report the number of files indexed, chunk count, and overall status.

index

Scanned

npx machina-cli add skill eidetics/claude-eidetic/index --openclaw

Files (1)

SKILL.md

993 B

/index

Usage:

/eidetic:index — index current project
/eidetic:index /path/to/project — index specific path

Step 1: Detect Path

Use the argument if provided. Otherwise:

git rev-parse --show-toplevel 2>/dev/null || echo "NO_GIT_REPO"

If NO_GIT_REPO and no argument, ask for path. Store as PROJECT_PATH.

Step 2: Dry Run

index_codebase(path="<PROJECT_PATH>", dryRun=true)

Show file count, extensions, top directories, warnings, and estimated cost if available.

Step 3: Index

index_codebase(path="<PROJECT_PATH>")

Step 4: Verify

get_indexing_status(path="<PROJECT_PATH>")

Report: files indexed, chunk count, status.

Step 5: What's Next

Suggest these next actions:

search_code("how does X work") — try a semantic search on this codebase
/cache-docs <library> — cache docs for a library you use frequently (e.g., React, Express)

Source

git clone https://github.com/eidetics/claude-eidetic/blob/main/plugin/plugins/claude-eidetic/skills/index/SKILL.mdView on GitHub

Overview

This skill indexes a codebase to enable Eidetic semantic search across code. It guides you from path detection to a full index, with a dry run to preview scope and cost. Once indexed, you can verify progress and perform semantic queries like 'how does X work'.

How This Skill Works

It starts by detecting the project path from an argument or by inferring the git root. It can perform a dry run via index_codebase with dryRun=true to show file counts, extensions, top directories, warnings, and estimated cost. Then it runs index_codebase without dryRun and finally uses get_indexing_status to report files indexed, chunk count, and status.

When to Use It

You want to index the current repository with /eidetic:index
You need to index a specific directory or path for targeted search
You want a non-destructive dry run to preview scope and cost before indexing
You want to confirm indexing completion and status via get_indexing_status
You want to perform follow-up actions like semantic search or caching docs after indexing

Quick Start

Step 1: Detect the project path using the provided argument or git rev-parse to determine PROJECT_PATH
Step 2: Dry run with index_codebase(path=PROJECT_PATH, dryRun=true) to preview files, extensions, and cost
Step 3: Run the full index with index_codebase(path=PROJECT_PATH) and then verify with get_indexing_status(path=PROJECT_PATH)

Best Practices

Prefer a dry run before large repos to estimate scope and cost
Ensure you have a clear PROJECT_PATH when indexing
Review file counts, extensions, and top directories surfaced by the dry run
Monitor the chunk count to manage index size and performance
After indexing, use search_code and cache-docs to maximize value

Example Use Cases

Index the current project and run a semantic search for 'how does X work'
Index a specific subfolder, e.g., services/auth, to focus search results
Dry run a monorepo to estimate file counts and potential costs before indexing
Index a repo, then verify with get_indexing_status to confirm completion
Post-index, cache docs for popular libraries (e.g., React, Express)

Frequently Asked Questions

Add this skill to your agents