- Home
- Skills
- Testing & Debugging
- reverse engineering kit
Works with the AI Agents you already use
reverse engineering kit
An evidence-based reverse engineering kit for verifiable analysis of binaries, APKs, and network protocols.
Free
reverse engineering kit
Example session with this skill installed
I need to analyze this Android APK to find where it sends user data. I have authorization for this research. Run the initial triage.
- Read your context and instructions
- Compiled the reverse engineering kit
I have initialized the analysis environment and performed the safety check. The authorization details have been recorded in notes.md, and the target hash is logged.
Triage Results: Android Analysis
| Field | Value |
|---|---|
| File | original/app-release.apk |
| Family | android |
| Package | com.example.tracker |
| Main Activity | com.example.tracker.MainActivity |
| Signals | okhttp3, firebase-analytics, custom-obfuscation |
| Read Next | references/android.md, references/anti-analysis.md |
The triage identified OkHttp3 and Firebase signatures. I have created the re-work/app-release/ directory structure. Initial impression: The app uses standard networking libraries but shows signals of custom obfuscation in the lib/ directory.
Next steps
- Run
scripts/decompile_apk.pyto extract the manifest and Java sources. - Execute
scripts/find_endpoints.pyto identify hardcoded URLs and API paths. - Formulate a hypothesis regarding the data exfiltration trigger in
notes.md.
Connects securely to your tools. The creator never sees your data.
About this skill
Software analysis often relies on guesswork and unverified decompilation. This skill provides an evidence-first framework for reverse engineering binaries, mobile apps, and network protocols, ensuring every claim is backed by reproducible tests.
What it does
- Automated Triage identifies file types, headers, and potential packers for native, Android, and JS targets.
- Ledger Management tracks findings across four confidence levels, from initial inference to verified proof.
- Multi-Target Analysis drives Ghidra for native code, jadx for APKs, and specialized tools for Electron or minified JS.
- Evidence Verification runs Python-based differential tests to confirm that recovered functions or algorithms match original behavior.
- Safety Enforcement implements strict gates for authorization, environment isolation, and data redaction.
How it works
- Safety and Setup: Run
doctor.pyto check tool chains, establish authorization innotes.md, and hash original files. - Triage and Route: Execute
triage.pyto determine the binary family and select the appropriate playbook for analysis. - Hypothesis Loop: Add claims to the ledger using
ledger.py, moving findings frominferredtoobservedvia tool output. - Verification: Write Python checks in
tests/and useledger.py verifyto promote findings toconfirmedstatus. - Reporting: Generate a structured
REPORT.mdusingmake_report.pyonce all mandatory sections and evidence are logged.
Frameworks & tools
The kit integrates with Ghidra (headless), LIEF, rizin, jadx, apktool, webcrack, and Frida. It requires Python 3.10+ for the orchestration scripts and ledger system.
Why this beats prompting it yourself
Standard prompts often hallucinate logic from decompiled code without verification. This skill enforces a test-driven workflow where claims must survive differential analysis and automated auditing before being accepted.
Use cases
- Recovering source logic from minified JavaScript or obfuscated Electron apps.
- Analyzing Android APKs to map API endpoints and internal data structures.
- Decoding proprietary network protocols from pcap/pcapng captures.
- Solving complex CTF challenges or crackmes with a formal evidence ledger.
Known limitations
Does not support .NET, Java JARs, Python bytecode (pyc), or iOS IPA files in the current version. Requires a pre-configured VM for dynamic analysis and script-based verification.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Testing & Debugging.
No support for .NET, JAR, pyc, or iOS. Requires external tools like Ghidra/Jadx. Dynamic analysis requires a user-provided VM.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Free to use
No account required to browse
Trust & safety
Security scan
Security scan passed 5 days ago
- Free with an Agensi account
- Software
PythonClaude CodeGhidra 12.2JDK 25rizinLIEFUPX 5.3jadx 1.5.6Node 22 or 24Kaitai Struct compiler 0.11Frida 17.25Ida Pro MCP
Git
uv (Python tool runner)
Install only the tools for the targets you analyze. python scripts/doctor.py --for native (or android, js, protocols, dynamic) shows what is missing and prints the install command for your system. It never installs anything itself. On Apple Silicon and ARM Linux, build Ghidra's decompiler once with