More screenshots
Works with the AI Agents you already use
.NET Reversing: Trust the IL, Not the Decompiler
Reverse-engineer .NET and C# binaries — confirm managed, deobfuscate on a copy, decompile, verify in IL, extract configs and C2 logic, and patch at the IL level with minimal, re-verified edits.
$5
.NET Reversing: Trust the IL, Not the Decompiler
Example session with this skill installed
I have a .NET info-stealer sample, stealer.exe. Strings are garbage and type names are single characters. Extract its C2 config and, if the task allows, patch out the exfiltration.
- Read your context and instructions
- Compiled the net reversing
Analysis — stealer.exe (managed, ConfuserEx, config extracted, patch verified)
Step 1 — confirmed managed: 64-bit PE, CLR header present, #Strings stream visible. Not native, not IL2CPP — the .NET playbook applies.
Step 2 — deobfuscate: packer detector names ConfuserEx. de4dot run on a copy (original untouched) produces stealer-clean.exe. Control-flow obfuscation is gone; string encryption survived — noted, not ignored. The string decryptor is one centralized method.
Step 3 — decompile, then verify in IL: ilspycmd on the clean copy gives readable C# for orientation. The C# view shows a simple license-check comparison; the IL shows an additional null check the decompiler smoothed away. Recorded — the IL wins, and the patch targets the IL, not the C# view.
Step 4 — extract: the decryptor is a single XOR routine with a 4-byte key. Reimplementing it decrypts the whole string table at once: C2 URL, 300-second exfiltration interval, and the targeted browser paths. All saved to the case directory with the routine's offset and the key, so another analyst can re-derive them.
Step 5 — patch: one IL instruction flipped (brfalse → brtrue) at the exfiltration gate. Re-verified: the assembly loads, metadata is consistent, and the specific behavior — no exfiltration on launch — is confirmed. The patched binary is a separate file; original and clean copy untouched. Rubric score: 8/8.
Connects securely to your tools. The creator never sees your data.
About this skill
.NET binaries are unusually kind to reversers: metadata, type names, and often string literals ship inside the assembly. This skill turns that kindness into a disciplined workflow: confirm the target is actually managed (CLR header, mscoree.dll, metadata streams — and the IL2CPP trap that makes Unity games native despite C# origins), strip obfuscation with de4dot on a copy before reading anything, decompile for orientation with dnSpyEx or ilspycmd, then switch to the IL view for anything that matters — because the C# decompiler smooths state machines and can misrender the exact logic you need. The IL is the ground truth.
Extraction targets the usual prizes: C2 URLs, keys, intervals (often runtime-decrypted — find the routine and reimplement it), persistence, and the centralized string decryptor that breaks everything at once. Patching happens at the IL level with minimal edits, re-verified after. A worked example takes an info-stealer from ConfuserEx obfuscation to a decrypted string table and a one-instruction patch. Red-team samples get diffed against public source first — the delta is the finding. A rubric gates the analysis: below 6 out of 8, the findings are not defensible yet.
- It is .NET-only: native binaries and IL2CPP Unity games belong in a native playbook, and the skill's first step exists to make that call before any time is spent.
- Deobfuscation depends on the obfuscator: de4dot handles the common ones, but surviving protections (string encryption, anti-tamper) need specific handling, not blind re-runs.
- The C# decompiler can misrender state machines and comparisons — which is why the IL is the ground truth and why patching from the C# view is forbidden by the skill.
- Recompiling from decompiled C# is not a substitute for IL patching: it produces a new binary with new metadata and new bugs.
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 any payment.
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
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 3 days ago