Skills
skills/affaan-m/ecc/regex-vs-llm-structured-text/Gen Agent Trust Hub

regex-vs-llm-structured-text

Pass

Audited by Gen Agent Trust Hub on May 22, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The validate_with_llm function in SKILL.md interpolates untrusted input (original_text) directly into an LLM prompt template. This creates a surface for indirect prompt injection where malicious instructions embedded in the text being parsed could influence the LLM's validation output.
  • Ingestion points: content parameter in process_document and original_text in validate_with_llm.
  • Boundary markers: Absent; the user input is concatenated with the prompt instructions.
  • Capability inventory: Invokes an LLM (claude-haiku-4-5-20251001) via a message client.
  • Sanitization: No sanitization or escaping is performed on the input text before prompt interpolation.
Audit Metadata
Risk Level
SAFE
Analyzed
May 22, 2026, 02:06 PM
Security Audit — agent-trust-hub — regex-vs-llm-structured-text