Help us improve
Share bugs, ideas, or general feedback.
Share bugs, ideas, or general feedback.
Share bugs, ideas, or general feedback.
By ahostbr
Human-in-the-loop result filtering. Filter search results BEFORE they pollute your context.
npx claudepluginhub ahostbr/claude-interactiveHuman-in-the-loop result filtering for Claude Code. Filter search results BEFORE they pollute your context window.
Every search result returned by Grep, Glob, or WebSearch consumes tokens in your context window. Large codebases can return dozens of matches, many irrelevant. By the time Claude processes them, your context is full of noise.
Claude Interactive lets you select which search results to keep before they enter context:
Copy this folder to your Claude Code plugins directory:
# macOS/Linux
cp -r claude-interactive ~/.claude/plugins/
# Windows (PowerShell)
Copy-Item -Recurse claude-interactive "$env:USERPROFILE\.claude\plugins\"
Restart Claude Code.
Add trigger words to your search requests:
"Find all TODO comments interactively"
"Search for useState and let me pick which files"
"Grep for errors, filter results first"
Trigger phrases:
interactivelylet me pickfilter resultsselect resultsEnable auto-mode to make ALL searches interactive:
/interactive auto on # Enable
/interactive auto off # Disable
/interactive status # Check status
When auto-mode is enabled, every Grep, Glob, and WebSearch will prompt for result selection.
You: "Find files with useState interactively"
Claude:
Found 5 results for "useState":
1. src/App.tsx:12 - const [count, setCount] = useState(0)
2. src/Form.tsx:8 - const [value, setValue] = useState('')
3. src/Modal.tsx:15 - const [open, setOpen] = useState(false)
4. src/List.tsx:22 - const [items, setItems] = useState([])
5. src/tests/App.test.tsx:5 - import { useState } from 'react'
Select which results to keep: [1] [2] [3] [All] [None]
You: Select 1, 2, 3
Claude: (continues with only App.tsx, Form.tsx, Modal.tsx)
This plugin uses Claude Code's native features:
.claude-interactive-automode for auto-mode toggleNo external dependencies. No MCP servers. No GUI required.
Context is precious. In a 200k token window:
A single broad Grep can return 50+ matches, each consuming tokens. With interactive filtering, you keep only what matters.
MIT
Share bugs, ideas, or general feedback.
Based on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
Open-source, local-first Claude Code plugin for token reduction, context compression, and cost optimization using hybrid RAG retrieval (BM25 + vector search), reranking, AST-aware chunking, and compact context packets.
Optimized file search, semantic indexing, and persistent memory for Claude Code — with optional sync to a self-hosted web dashboard
Claude Code context optimizer. Builds a local static repo graph, injects ranked file:line candidates through a UserPromptSubmit hook, and adds a SessionStart prewarm brief. No embeddings, no server, no telemetry. Measured live A/B: 40.9% aggregate token reduction across 36 Claude Code calls.
Fast and token-friendly code reading for AI coding agents. Symbol-aware MCP tools that replace cat/grep with ~85% fewer tokens, sub-millisecond search, and a raw fallback that preserves cat/grep parity byte-for-byte.
Semantic code search powered by ColBERT. Replaces grep/ripgrep with natural language understanding for smarter code navigation.
Enrich coding agent prompts with codebase context using a local LLM
Automatic transcript backup. Every conversation saved, organized by date, fully searchable.
Multi-CLI agent orchestration — spawn, message, and coordinate AI coding agents across Claude Code, Codex, Copilot, and more from a single harness. 5-tier hierarchy, inter-agent messaging, polymathic thinking agents, git-as-memory.
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge.
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge.
Sign in to claim