evasion_integrate — Claude Code command: what it does, how to install
evasion_integrate adds evasion techniques to a shellcode loader you already have: it edits the source, checks which tricks fit and builds the binary.
MIT ★ 336 updated 2026-04-09
.claude/commands/evasion_integrate.mdWhat evasion_integrate does
The job goes to the evasion-agent subagent. It opens the loader source at the path you give, pulls matching techniques from its knowledge base, tests them against the code and rewrites the file. A compile follows, then a report on the techniques applied, the APIs they touched and the detection risk.
You can let the command choose or narrow the set: --type for a category such as api_obfuscation or execution_evasion, --complexity to filter by how hard a technique is, --technique for specific IDs. Results land in output/evasion_<id>.c and output/evasion_<id>.exe.
The loader path is not optional, so put it first and the filters after; without it there is nothing to read. The compiled executable is never launched. A clean build is where the run ends.
What it helps with
- add API hashing to an existing loader
- encrypt strings without rewriting the loader by hand
- switch API calls to direct or indirect syscalls
- check which evasion techniques fit the current source
- compile the patched loader and get a list of changes
Example prompts for the agent
- Integrate evasion into this loader and pick the techniques yourself:
/evasion_integrate /path/to/loader.c - Only string techniques here:
/evasion_integrate /path/to/loader.c --type string_obfuscation - Keep it to the simple techniques:
/evasion_integrate /path/to/loader.c --complexity simple
How to install evasion_integrate
Manually: put the files in place
.claude/commands/evasion_integrate.mdCall it with /evasion_integrate. For all projects use ~/.claude/commands/
Manually: put the files in place
~/.codex/prompts/evasion_integrate.mdCall it with /prompts:evasion_integrate. Codex marks this format deprecated in favour of skills
Manually: put the files in place
.github/prompts/evasion_integrate.prompt.mdManually: put the files in place
.gemini/commands/evasion_integrate.tomlGemini stores commands as TOML: the prompt text goes into the prompt field
Questions about evasion_integrate
How do I install evasion_integrate in Claude Code?
Copy the command file to .claude/commands/evasion_integrate.md in your project. /evasion_integrate is then available in Claude.
What does evasion_integrate do?
It spawns the evasion-agent subagent, which reads your loader source, selects evasion techniques that fit, applies them, compiles the result and lists every change.
Does it run the compiled executable?
No. It only modifies the source you provide and stops once the build succeeds. The generated .exe stays in output/.
Source text
evasion_integrate.md · MIT
--- description: Integrate evasion techniques into existing shellcode loaders. Add API hashing, string encryption, syscalls, etc. to user-provided code. Triggers on "integrate evasion", "add evasion", "bypass", "hardening", "免杀", "二开". argument-hint: Required: path to loader source (e.g., "/path/to/loader.c") + optional technique filters --- # Evasion Integrate Command Launch the evasion-agent to add evasion techniques to an existing loader. ## Usage ```bash /evasion_integrate /path/to/loader.c # Auto-select techniques /evasion_integrate /path/to/loader.c --type api_obfuscation # Specific type /evasion_integrate /path/to/loader.c --type string_obfuscation,anti_analysis /evasion_integrate /path/to/loader.c --complexity simple # Filter by complexity /evasion_integrate /path/to/loader.c --technique T001,T003 # Specific IDs ``` ## What This Command Does 1. **Reads** user-provided loader source code 2. **Queries** evasion techniques from knowledge base 3. **Analyzes** compatibility with the loader 4. **Integrates** selected techniques into the code 5. **Compiles** modified loader 6. **Reports** all changes made ## Evasion Types | Type | Description | Complexity | |------|-------------|------------| | api_obfuscation | API hashing, PEB walking | medium | | string_obfuscation | XOR encryption, stack strings | simple | | memory_evasion | Permission flipping (RW→RX) | simple | | execution_evasion | Direct/indirect syscall | complex | | anti_analysis | Anti-debug, anti-VM | medium | | amsi_etw_bypass | AMSI/ETW patching | medium | | unhooking | NTDLL unhooking | complex | ## Output - Modified source: `output/evasion_<id>.c` - Compiled executable: `output/evasion_<id>.exe` - Changes summary: - Techniques applied - APIs modified - Detection risk assessment ## Security - **ONLY** modify user-provided code - **NEVER** run generated executables - Compilation success is sufficient ## Agent Spawns `evasion-agent` subagent for technique integration. See `evasion_integrate` skill for detailed patterns.
© JDArmy · MIT · JDArmy/Evasion-SubAgents