regex-vs-llm-structured-text
Pass
Audited by Gen Agent Trust Hub on Mar 26, 2026
Risk Level: SAFE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external text content through an LLM validator, creating a surface for indirect prompt injection.
- Ingestion points: The
contentparameter inprocess_documentand theoriginal_textparameter invalidate_with_llm(SKILL.md). - Boundary markers: The LLM prompt uses basic string interpolation (
Text: {original_text}) without dedicated boundary markers (like XML tags or specific delimiters) to isolate potentially untrusted input from instructions. - Capability inventory: The skill utilizes an LLM client (
client.messages.create) to process and potentially 'correct' data based on the input text. - Sanitization: No explicit sanitization or filtering of the input text is performed before it is sent to the LLM.
Audit Metadata