context-bench

Your AI editor knows what you're working on — automatically.

A self-learning Claude Code plugin that detects your current project from your first message and injects the right file context — without any manual configuration.

How it works

1. You type your first message
2. context-bench reads the prompt (UserPromptSubmit hook)
3. Keywords are matched against your project topics
4. Relevant files are injected as invisible context
5. Claude responds with full project context already loaded
6. File changes are tracked (PostToolUse hook)
7. Confidence scores update at session end (SessionEnd hook)

Installation

As Claude Code Plugin (recommended)

claude plugin add nessos666/context-bench

That's it. The plugin registers all hooks automatically.

Standalone (manual)

git clone https://github.com/nessos666/context-bench
cd context-bench
./install.sh

install.sh registers three hooks in ~/.claude/settings.json and copies context_bench.py to ~/.context-bench/.

Local testing

claude --plugin-dir ./context-bench

Toggle (disable/enable)

To temporarily disable context-bench without uninstalling:

touch ~/.context-bench/DISABLED   # disable
rm ~/.context-bench/DISABLED      # enable again

When disabled, all hooks exit immediately (no context injected, no tracking). The learn hook still cleans up session files to prevent leaks.

If you use Claude Code skills, you can add shortcuts:

~/.claude/commands/ctx-bench-aus.md:

Run: mkdir -p ~/.context-bench && touch ~/.context-bench/DISABLED
Output: "context-bench deactivated"

~/.claude/commands/ctx-bench-an.md:

Run: rm -f ~/.context-bench/DISABLED
Output: "context-bench active"

Uninstall

Plugin mode:

claude plugin remove context-bench

Standalone:

./uninstall.sh

How it learns

Each project topic has a confidence score (0–1):

• Match found + files changed → +0.15
• Match found + no files changed → −0.05
• No match + files changed → new topic created with 0.5
• Topic idle > 30 days → gradual decay (−0.01/day)
• Topic below 0.3 → removed automatically

Topics are stored in ~/.context-bench/projects.json.

Bootstrap

On first run (no projects.json yet), context-bench scans the current directory for framework markers:

Marker	Detected as
pyproject.toml / requirements.txt / setup.py	Python project
package.json	Node.js project
go.mod	Go project
Cargo.toml	Rust project
pom.xml	Java project

The bootstrap completes within 200ms to avoid blocking hooks.

Architecture

context_bench.py (single file, zero dependencies)
├── cmd_prompt()  → UserPromptSubmit  → match topic, inject context
├── cmd_track()   → PostToolUse       → record changed files
└── cmd_learn()   → SessionEnd        → update confidence, create topics, decay

Data model:

• Topic: id, keywords, root, paths, confidence, uses, last_used
• Database: version, projects[], settings
• Session: matched_topic, changed_files, prompt, injected_paths, cwd

Requirements

• Python 3.9+ (stdlib only, no dependencies)
• Claude Code with hooks support

License

MIT