bilingual-source-archive
Pass
Audited by Gen Agent Trust Hub on Oct 30, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from web articles, pasted text, and browser captures, creating an attack surface for indirect prompt injection where malicious instructions could be embedded in the source text.
- Ingestion points: Source text is acquired from user-provided attachments, pasted text, public URLs, or existing platform pages as defined in the 'Source Acquisition' section of SKILL.md.
- Boundary markers: The instructions do not define explicit delimiters or 'ignore' commands to prevent the agent from executing instructions found within the source content.
- Capability inventory: The skill leverages platform-specific workflows (Feishu, Notion, Obsidian) to write and update document content.
- Sanitization: The included validation script (validate_bilingual_source.py) focuses on structural linting and formatting residue but does not filter for malicious instruction patterns.
- [COMMAND_EXECUTION]: The skill includes instructions to execute a local Python script to validate Markdown drafts before finalization.
- Evidence: SKILL.md contains a validation step calling
python .tools/skills/bilingual-source-archive/scripts/validate_bilingual_source.py. The script performs regex-based checks on the content for forbidden labels and residue.
Audit Metadata