wycheproof
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill provides instructions to fetch cryptographic test vectors from the Project Wycheproof repository maintained by C2SP. \n
- Evidence in
SKILL.md: The skill referenceshttps://raw.githubusercontent.com/C2SP/wycheproof/master/testvectors_v1/for fetching JSON files and suggests addinghttps://github.com/C2SP/wycheproof.gitas a git submodule.\n- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process JSON test vectors from external sources, which represents a potential attack surface.\n - Ingestion points: Test data is loaded from files in
SKILL.mdviajson.loadsandfs.readFilefollowing a download from the C2SP repository.\n - Boundary markers: The implementation examples do not use specific delimiters or instructions to ignore potential instructions embedded in the test data.\n
- Capability inventory: The skill demonstrates the use of
curlfor network data acquisition and standard file system access for reading the test vectors.\n - Sanitization: The provided code examples focus on cryptographic validation and do not include specific sanitization for prompt injection in the ingested JSON strings.
Audit Metadata