Codex/jsconfuser vm two step decoder

.gitignore

@@ -107,3 +107,7 @@ input.js
 output.js
 .vscode/
 .idea/
+
+# Qualification workspace: encoder checkouts fetched by scripts/qualify-jsconfuser-vm
+sandbox-tests/
+__pycache__/

README.md

@@ -7,12 +7,15 @@ readable source.
 
 One target per run, selected with `-t`:
 
+The result-aware `jsconfuser-vm` target is listed in the dedicated section below.
+
 | `-t` | target |
 |---|---|
 | `common` | frequently-seen local obfuscation, not tied to any one tool — unreachable code, nested blocks, constant expressions, raw strings |
 | `jjencode` | jjencode, in the variant emitted by sojson.com |
 | `sojson` | sojson |
 | `sojsonv7` | sojson v7 |
+| `jsconfuser-vm` | JS-Confuser numeric VM containers via the standalone result-aware adapter |
 | `obfuscator` | [javascript-obfuscator](https://github.com/javascript-obfuscator/javascript-obfuscator) (obfuscator.io) |
 | `obfuscatorx` | the same obfuscator, version-aware — see below for how it differs |
 | `jsconfuser` | [JS-Confuser](https://github.com/MichaelXF/js-confuser) |
@@ -47,7 +50,17 @@ A sample whose range overlaps that gap is reported as unknown rather than treate
 ### `jsconfuser`
 
 Covers JS-Confuser 2.x up to and including the `high` preset. Which transforms are
-reversed, which are not, and why: [docs/jsconfuser.md](docs/jsconfuser.md).
+reversed, which are not, and why are documented in the
+[decode-nexus jsconfuser plugin reference](https://github.com/echo094/decode-nexus/blob/main/skills/decode-js/plugins/jsconfuser.md).
+
+### `jsconfuser-vm`
+
+Decodes accepted numeric VM containers without executing target code. Each attempt writes the
+adapter's JavaScript output and a JSON result record containing `status`, `diagnostic`, and the
+standalone schema/proof metadata. The result path defaults to `<output>.result.json`; pass
+`--result path.json` to choose an explicit path. A declined input is written byte-for-byte to the
+output, recorded with `status: "declined"`, and exits successfully. The `--result` option is
+target-specific; legacy targets keep their string/null output contract and do not write sidecars.
 
 ## Usage
 
@@ -71,7 +84,7 @@ npm run decode -- -t type [-i input.js] [-o output.js] [-v]
 ```
 
 `xxx` is one of the predefined commands, each a shorthand for one target — `deob`,
-`dejsc`, `deso`, `desov7`. See the `scripts` field in [package.json](package.json).
+`dejsc`, `dejsc-vm`, `deso`, `desov7`. See the `scripts` field in [package.json](package.json).
 
 The default input file is `input.js`. The file cannot contain additional codes other
 than obfuscated code (such as non-obfuscated code).
@@ -80,6 +93,12 @@ The default output file is `output.js`.
 
 `-v` turns on per-pass progress tracing, which is off by default.
 
+For the result-aware VM target, use `--result` when an explicit result path is preferred:
+
+```shell
+npm run decode -- -t jsconfuser-vm -i encoded.js -o decoded.js --result decoded.json
+```
+
 ## Related Projects
 
 * [cilame/v_jstools](https://github.com/cilame/v_jstools)

package.json

@@ -10,9 +10,11 @@
     "deob": "node src/main.js -t obfuscator",
     "deobx": "node src/main.js -t obfuscatorx",
     "dejsc": "node src/main.js -t jsconfuser",
+    "dejsc-vm": "node src/main.js -t jsconfuser-vm",
     "deso": "node src/main.js -t sojson",
     "desov7": "node src/main.js -t sojsonv7",
     "test": "vitest --config vitest.config.js",
+    "test:light": "vitest --config vitest.light.config.js",
     "lint": "eslint --fix src"
   },
   "dependencies": {