ccf-submission-checker
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content from user-provided manuscripts, LaTeX source files, and artifact READMEs during the checking process.
- Ingestion points: SKILL.md workflow steps 2, 4, and 5 specify reading ccfa.yaml, TeX/PDF files, and artifact files.
- Boundary markers: The skill does not define specific delimiters to separate paper content from agent instructions, though it focuses on metadata and formatting rather than content interpretation.
- Capability inventory: The skill identifies venue rules and checks file structure but does not include scripts that execute arbitrary shell commands or perform network operations.
- Sanitization: No specific sanitization of input text is mentioned, but the prompt explicitly restricts the agent from submitting or uploading the paper.
Audit Metadata