
from grapefruit1,316
CLI skill to control the IGF (Grapefruit) dynamic instrumentation server for mobile security testing: enumerate Frida devices, inspect apps, run hooks, access f
The IGF (Grapefruit) CLI skill lets an agent control a running IGF dynamic instrumentation server (default http://localhost:31337) to perform mobile app security analysis. It exposes REST and Socket.IO RPC operations to enumerate Frida devices, list and inspect apps/processes, query hook and network/history logs, manage files on device, dump memory, and run introspection (class lists, symbols, sqlite). Typical workflows include triaging running processes, extracting logs, running hooks and pins (sslpinning, crypto), and gathering artifacts for vulnerability analysis.
Use this skill when interacting with a local or remote IGF/Grapefruit server during mobile security testing: to list connected Frida devices, attach to an app, run hooks, inspect classes or modules, read logs, retrieve files, or perform binary/security checks. It's intended for security testers, reverse engineers, and developers needing deep runtime inspection.
Likely used by CLI-capable agents and code-oriented assistants (Copilot-style or GPT/Codex integrations) that can invoke shell commands or format JSON summaries of results.
IGF (Grapefruit) is a CLI skill for controlling a dynamic instrumentation server used in mobile security testing via Frida. The SKILL.md is comprehensive — well-structured command reference covering device management, hooking, memory inspection, class dumping, and platform-specific features for Android/iOS. No bundled scripts; the skill instructs the agent to call the `igf` CLI directly. No security issues found: no credential leakage, no destructive unguarded commands, no data exfiltration. The eval command carries inherent risk but is documented as a feature. Niche audience — mobile security researchers and reverse engineers.
Well-documented CLI reference skill. The `igf agent eval` and `igf agent memory dump` commands are powerful but expected for a security testing tool. No exfiltration or self-updating mechanisms. The skill is purely a command reference — no scripts, no network calls, no hidden behavior. Usefulness is moderate due to niche audience (mobile security testers).