Academic Figure Prompt

Pass

Audited by Gen Agent Trust Hub on Oct 27, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill instructions require the agent to read and analyze user-provided document files (LaTeX, Word, PDF) to extract information for prompt generation, creating a surface for indirect prompt injection.
  • Ingestion points: User-supplied document files (LaTeX, Word, PDF) as described in "Step 1: 理解论文内容" in SKILL.md.
  • Boundary markers: Absent. There are no instructions to use delimiters or ignore potentially malicious instructions embedded within the processed documents.
  • Capability inventory: Primary capability is text generation (image prompts). While "bash" is listed in the allowed tools in the frontmatter, no explicit shell commands are provided in the skill body for data manipulation.
  • Sanitization: Absent. The skill does not provide methods for filtering or sanitizing the content extracted from external documents.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 27, 2026, 04:44 AM