atcoder-client
Pass
Audited by Gen Agent Trust Hub on Nov 13, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data from the AtCoder platform and AtCoder Problems API.
- Ingestion points: The skill fetches contest problems, submissions, and metadata from atcoder.jp and kenkoooo.com (SKILL.md).
- Boundary markers: No delimiters or specific instructions to ignore embedded commands within fetched data are provided.
- Capability inventory: The skill allows the use of powerful tools such as Bash, Write, and WebFetch (SKILL.md).
- Sanitization: No sanitization or validation of external data is documented before processing.
- [EXTERNAL_DOWNLOADS]: The skill fetches data from well-known competitive programming services.
- Evidence: Integrates with the AtCoder web interface and the AtCoder Problems (kenkoooo.com) API to retrieve contest and problem metadata.
Audit Metadata