Frida AI Skills

Model-agnostic AI skills for planning, writing, packaging, and debugging Frida instrumentation across Android, iOS, native libraries, desktop processes, and Frida Gadget workflows.

Tags: ai-skills, agent-skills, codex-skills, frida, reverse-engineering, dynamic-instrumentation, android, ios, native-hooks, mobile-security, appsec, ssl-pinning, frida-gadget

AI Agent Compatibility

These skills are written as plain SKILL.md instruction folders so any AI agent or skill runner can use them by reading the relevant skill directory. The agents/openai.yaml files are compatibility metadata for OpenAI/Codex-style skill lists; they do not make the skill content OpenAI-only.

Install With npx

List available skills without installing:

npx skills add Rudra-ravi/frida-skills --list --full-depth

Install every skill from this repository for Codex:

npx skills add Rudra-ravi/frida-skills --skill '*' --agent codex --yes --full-depth

Install every skill globally for every supported agent:

npx skills add Rudra-ravi/frida-skills --all --global --yes --full-depth

Install to selected agents:

npx skills add Rudra-ravi/frida-skills --skill '*' --agent codex --agent claude-code --agent cursor --yes --full-depth

Install one focused skill:

npx skills add Rudra-ravi/frida-skills --skill frida-workflow --agent codex --yes --full-depth

Install from the full GitHub URL:

npx skills add https://github.com/Rudra-ravi/frida-skills --skill frida-tls-pinning --agent codex --yes --full-depth

Install from a direct skill folder URL:

npx skills add https://github.com/Rudra-ravi/frida-skills/tree/main/frida-codeshare-search --agent codex --yes --full-depth

Use --copy if your agent setup cannot use symlinks:

npx skills add Rudra-ravi/frida-skills --skill '*' --agent codex --copy --yes --full-depth

Restart your AI agent after installing so the new skills are discovered.

For non-Codex agents, clone this repository and register or copy the skill folders your agent supports. Each skill is self-contained in its own directory and uses the same required structure:

skill-name/
├── SKILL.md
├── agents/openai.yaml
├── references/
└── scripts/

Codex/OpenAI-Compatible Skill Installer

You can also install with an OpenAI/Codex-compatible skill installer from the GitHub URL:

$skill-installer install https://github.com/Rudra-ravi/frida-skills/tree/main/frida-workflow

Repeat with any skill folder below.

Skills

Skill	Use When
frida-workflow	Planning Frida tasks, selecting attach/spawn/Gadget mode, chaining focused skills, and preserving reproducible evidence.
frida-setup-checklist	Verifying host tools, Android/iOS devices, transports, Frida versions, architecture, and smoke scripts before hook work.
frida-tools-reference	Choosing and running Frida CLI tools such as frida, frida-ps, frida-ls-devices, frida-trace, frida-discover, frida-kill, and frida-create.
frida-tracing-discovery	Discovering classes, methods, modules, exports, imports, symbols, stack traces, and call paths before writing final hooks.
frida-native-hooks	Hooking native exports, imports, stripped functions, memory, Interceptor, NativeFunction, NativeCallback, Stalker, and CModule paths.
frida-android-hooks	Instrumenting Android Java/Kotlin/JNI behavior, overloads, constructors, class loaders, pinning, and native library transitions.
frida-ios-hooks	Instrumenting Objective-C, Swift/native, iOS trust paths, rootless jailbreak setups, and Gadget workflows.
frida-agent-builder	Creating maintainable TypeScript Frida agents with structured logging, RPC probes, and multi-hook organization.
frida-gadget-injection	Planning and debugging Frida Gadget workflows for Android, iOS, desktop apps, jailed devices, embedded configs, and load timing.
frida-tls-pinning	Identifying and hooking TLS/SSL pinning paths across Android, iOS, Flutter, React Native, native TLS stacks, and app-specific trust code.
frida-anti-instrumentation	Diagnosing and bypassing anti-Frida, anti-debugging, root/jailbreak, emulator, ptrace, port scan, module scan, and integrity checks.
frida-data-extraction	Observing runtime data such as HTTP requests, headers, tokens, JSON, crypto values, preferences, SQLite, files, and memory buffers.
frida-codeshare-search	Searching Frida CodeShare with curl, extracting scripts from project pages, and converting public snippets into reviewed local agents or --codeshare commands.
frida-script-review	Reviewing, hardening, and simplifying public, generated, broad, native-pointer-heavy, or mutation-heavy Frida scripts before running them.
frida-troubleshooting	Debugging attach/spawn failures, version mismatches, hooks that do not fire, runtime bridge issues, and target crashes.

Research Basis

The skills combine current official Frida documentation with practitioner patterns from Frida CodeShare, OWASP MASTG, and advanced Frida references. They emphasize observation before mutation, small reversible hooks, explicit version checks, and treating public bypass bundles as reconnaissance rather than final proof.

Primary sources are listed in frida-workflow/references/frida-sources.md.

Validate

for d in frida-workflow frida-setup-checklist frida-tools-reference frida-tracing-discovery frida-native-hooks frida-android-hooks frida-ios-hooks frida-agent-builder frida-gadget-injection frida-tls-pinning frida-anti-instrumentation frida-data-extraction frida-codeshare-search frida-script-review frida-troubleshooting; do
  python3 /home/rudra/.codex/skills/.system/skill-creator/scripts/quick_validate.py "$d"
done

bash -n frida-workflow/scripts/frida-env-check.sh
python3 -m py_compile frida-agent-builder/scripts/scaffold-frida-agent.py
python3 -m py_compile frida-codeshare-search/scripts/codeshare_extract.py
npx skills add . --list --full-depth

License

MIT. See LICENSE.