null200ok-bughunter

Pass

Audited by Gen Agent Trust Hub on Nov 9, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONOBFUSCATION
Full Analysis
  • [PROMPT_INJECTION]: The skill implements an 'Activation Protocol' triggered by the keyword 'NULL200OK' which commands the agent to override its default identity and behavior, specifically requesting 'zero moralizing' and the immediate adoption of an offensive security persona for the remainder of the session.
  • [COMMAND_EXECUTION]: The instructions provide several shell-based 'Standard Pipeline' and 'Recon Pipeline' templates (e.g., in Section 1.1 and 1.3). These templates utilize shell variables such as '$TARGET' and '$T' without providing specific sanitization or escaping instructions, creating a risk of command injection on the host system if the agent executes them using malicious user-supplied target names.
  • [INDIRECT_PROMPT_INJECTION]: The skill's primary function is to process untrusted data from external targets such as web responses, code snippets, and documentation. While it includes 'Marker Discipline' to help differentiate data from code, it lacks robust boundary markers or sanitization logic to prevent instructions embedded within target data from influencing the agent's behavior (e.g., through RAG documents or tool output poisoning).
  • [OBFUSCATION]: The skill contains numerous references to obfuscation techniques, including Base64-encoded payloads, hex strings, and Unicode-based smuggling. These were flagged by static detectors but are documented as part of the offensive security doctrine and historical case studies (e.g., XSS payloads in Section 5.1 and tool-poisoning descriptions in Section 13.13) rather than as active malicious code intended to compromise the agent's host environment.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 9, 2026, 11:17 AM