Why does this skill stop me from reading files?

It intercepts operations that would waste tokens or provide less relevant context than a ranked semantic search, suggesting a more efficient tool for the task.

Does it require claudemem to be installed?

Yes, it is designed to work with claudemem to perform semantic searches. It will even prompt you to index your project if it isn't already.

How much token usage can this save?

Reading five full files might cost 5,000+ tokens, whereas a targeted semantic search usually costs around 500 tokens while providing more relevant code chunks.

What specifically triggers the Search Interceptor?

The skill triggers when attempting to read 3 or more files at once, using broad glob patterns like **/*.ts, or planning sequential reads to 'understand' a concept.

Will it prevent me from reading a single specific file?

No. The interception decision matrix allows for 1-2 specific file reads or exact string matches via grep without redirecting you.

Semantic Search Interceptor

Name: Semantic Search Interceptor
Author: involvex

byinvolvex

Productivity & Workflow

Optimizes code exploration by intercepting bulk file operations and redirecting to efficient semantic search queries.

About

The Semantic Search Interceptor acts as an intelligent guardrail for Claude, preventing inefficient token usage and sequential file reading habits. By automatically triggering before bulk file reads or broad glob patterns, it evaluates if a semantic search via claudemem would provide better, more ranked context. This skill helps developers overcome tool familiarity bias, ensuring that the most relevant code sections are analyzed first while significantly reducing the overhead associated with reading entire directories or large files.

Key Features

Real-time validation of claudemem index status
Automatic interception of bulk file read operations (3+ files)
Intelligent redirection from sequential reads to ranked semantic queries
0 GitHub stars
Prevention of broad glob pattern execution to minimize token waste
Decision matrix for distinguishing between exact matches and semantic concepts

Use Cases

Investigating complex modules like authentication without reading every individual file in the directory
Discovering architectural patterns across a large codebase without performing expensive broad glob patterns
Auditing specific API integrations when multiple related paths are mentioned in a prompt

About

Key Features

Real-time validation of claudemem index status
Automatic interception of bulk file read operations (3+ files)
Intelligent redirection from sequential reads to ranked semantic queries
0 GitHub stars
Prevention of broad glob pattern execution to minimize token waste
Decision matrix for distinguishing between exact matches and semantic concepts

Use Cases

Investigating complex modules like authentication without reading every individual file in the directory
Discovering architectural patterns across a large codebase without performing expensive broad glob patterns
Auditing specific API integrations when multiple related paths are mentioned in a prompt