The Agent Skills Directory

[SAFE]: The skill is purely instructional and focused on security research and reverse engineering of obfuscated JavaScript runtimes. All scripts provided are intended for local execution by the user on target files.
[EXTERNAL_DOWNLOADS]: The skill recommends and provides snippets for well-known automation and testing frameworks.
Python: Uses selenium for browser automation and anti-detection testing.
Node.js: Uses playwright for intercepting network requests and interacting with DSL VM environments.
Context: These are standard libraries for the described purpose of dynamic analysis and do not involve unverified or risky third-party sources.
[PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted third-party code (target.js), creating a surface for indirect prompt injection.
Ingestion points: The Python snippets read the content of target.js for classification and opcode extraction.
Boundary markers: Not present in the provided analysis scripts; the agent relies on the user providing specific files for analysis.
Capability inventory: File reading (open), regex processing (re.findall), and browser automation (webdriver.Chrome).
Sanitization: Not present, as is typical for static analysis scripts designed to process raw code snippets.

dsl-vm-reverse