text-splitter

Pass

Audited by Gen Agent Trust Hub on Sep 14, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill is a local text processing tool with no external network dependencies or sensitive file access.\n- [SAFE]: No obfuscation, malicious persistence, or privilege escalation patterns were detected in the scripts or documentation.\n- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted text data for processing, which is a potential vector for indirect prompt injection. However, the risk is considered safe because the skill has no tool access (allowed-tools is empty) and performs no side effects like file writing or network requests.\n
  • Ingestion points: The text argument in scripts/text_splitter.py.\n
  • Boundary markers: The output report uses markdown headers but lacks strict delimiters for untrusted content.\n
  • Capability inventory: Limited to local string manipulation and regex operations; no subprocess or network tools.\n
  • Sanitization: No specific sanitization or filtering of input text is performed before processing or inclusion in output.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 14, 2026, 06:49 PM
Security Audit — agent-trust-hub — text-splitter