Quick Drop Mode

Quick Drop is a fast way to analyze files without writing a detailed prompt. Drop a file, add a brief instruction, and get results.

How It Works

1. Switch to Quick Drop tab in the side panel
2. Drop a file (or click to browse)
3. Add instruction (optional)
4. Click “Analyze”

Claude Code reads the file and responds based on your instruction (or provides a general analysis if no instruction given).

Supported File Types

Quick Drop works with any text-based file:

Category	Extensions
Code	.py, .r, .sh, .js, .ts, .go, .rs
Data	.csv, .tsv, .json, .yaml
Config	.toml, .ini, .conf
Documents	.md, .txt, .rst
Scripts	.sh, .bash, .zsh

Caution

Binary files (images, compiled executables, archives) are not supported. Use the Claude Code tab for tasks involving binary files on the server.

Example Use Cases

Code Review

Drop a Python file with instruction:

Review for bugs and performance issues

Claude analyzes the code and provides feedback.

Data Validation

Drop a CSV file with:

Check for missing values and outliers

Config Audit

Drop a config file with:

Security review - any sensitive values exposed?

Documentation

Drop a script with:

Generate docstrings for all functions

How It Differs from Claude Code Tab

Quick Drop	Claude Code Tab
Single file focus	Multi-file projects
Fast analysis	Full plan-execute workflow
Read-only by default	Can write files and run commands
No session continuity	Maintains context across tasks

Tip

Use Quick Drop for one-off analysis. Use Claude Code tab for tasks that modify files or require context from previous work.

File Size Limits

• Maximum file size: 1 MB
• For larger files, use the Claude Code tab with file paths

Response Format

Quick Drop responses appear directly in the panel:

📄 Analyzing: experiment.py (2.3 KB)

Analysis:

This script runs a machine learning experiment with
the following structure:

1. Data loading (lines 1-25)
   - Loads from data/train.csv
   - Applies StandardScaler normalization

2. Model definition (lines 27-45)
   - RandomForestClassifier with 100 estimators
   - Uses default hyperparameters

3. Training loop (lines 47-80)
   - 5-fold cross-validation
   - Tracks accuracy per fold

Suggestions:
- Add error handling for missing data file
- Consider hyperparameter tuning
- Log results to file instead of print()

---
Analyzed in 3.2s · Free (Max)

Permissions

Quick Drop uses Plan Only permissions by default. Claude Code can:

• ✅ Read the uploaded file
• ✅ Analyze and provide suggestions
• ❌ Write files
• ❌ Run commands

To enable execution (e.g., to apply suggested changes), use the Claude Code tab instead.

Tips

1. Be specific — “Check for SQL injection” is better than “Review security”
2. One file at a time — For multi-file analysis, use Claude Code tab
3. Use for iteration — Drop updated versions as you refine

Next Steps

• Activity Feed — Understanding Claude’s output
• Permission Profiles — Control what Claude can do