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 inreferences/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.mduses 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