subtitle-correction
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from user-provided subtitle files, creating a surface for indirect prompt injection.
- Ingestion points: User-uploaded .srt files are read and processed as described in
SKILL.mdand handled byscripts/subtitle_tool.py. - Boundary markers: The skill defines strict rules for timeline preservation and specific error categories (phonetic, technical, code) which serve as constraints on how the data is interpreted.
- Capability inventory: The skill utilizes file system access (read/write) and text processing via a Python script. It does not possess network or arbitrary shell execution capabilities.
- Sanitization: The
scripts/subtitle_tool.pyscript employshtml.escape()when generating HTML diff reports to sanitize subtitle content before display.
Audit Metadata