binary-diff

Pass

Audited by Gen Agent Trust Hub on Jul 4, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill exhibits a surface for Indirect Prompt Injection (Category 8) because it processes disassembly and decompiled code from external, potentially untrusted binaries.\n
  • Ingestion points: Untrusted data from binaries is interpolated into the {disasm_code} and {procedure} variables within the prompt template in references/prompt-template.md.\n
  • Boundary markers: The skill uses markdown code blocks (triple backticks) as delimiters to isolate binary-derived content from instructions in the LLM prompt.\n
  • Capability inventory: The logic in references/prompt-template.md uses the LLM output to perform function renaming (idapro_rename) and add comments (idapro_set_comments) in the IDA Pro environment.\n
  • Sanitization: The implementation correctly uses yaml.safe_load() in the provided Python script skeleton to prevent unsafe deserialization of the LLM response.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 4, 2026, 02:01 AM
Security Audit — agent-trust-hub — binary-diff