The Agent Skills Directory

[SAFE]: The skill consists of documentation and configuration files for planning and executing SumUp sandbox tests. A thorough analysis across all 11 threat categories revealed no malicious patterns or security risks.
[COMMAND_EXECUTION]: The skill includes a curl skeleton in SKILL.md for testing checkout creation. This is a documentation example that correctly uses a shell variable ($SUMUP_API_KEY) for authentication, which is the recommended safe practice for avoiding hardcoded secrets.
[DATA_EXFILTRATION]: The API endpoint referenced in the cURL example (api.sumup.com) is an official resource of the vendor 'sumup'. No patterns indicating unauthorized data access or transmission to suspicious domains were found.
[PROMPT_INJECTION]: The instructions in SKILL.md use standard instructional language for test planning and do not contain any markers or techniques designed to bypass AI safety guidelines or override system instructions.
[OBFUSCATION]: All analyzed files, including the SVG asset and YAML configuration, contain plain-text, human-readable content with no signs of Base64 encoding, hidden characters, or homoglyph attacks.

sumup-testing