anycap-worldcup-predict
Pass
Audited by Gen Agent Trust Hub on Jun 21, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill provides shell command templates for interacting with the anycap CLI. Examples in references/batch.md demonstrate how to use jq, sed, and shell loops to process multiple players. These are standard integration patterns for agents with shell access and are used here to structure and filter football data.
- [EXTERNAL_DOWNLOADS]: The skill retrieves data from the vendor's own services (anycap.ai) and well-known authoritative sources like Wikipedia (en.wikipedia.org). These network operations are intrinsic to the skill's primary function of providing up-to-date football information and verifying its accuracy.
- [PROMPT_INJECTION]: The instructions include specific guardrails to protect against behavioral manipulation. In references/predict.md, a mandatory disclaimer is required for all predictions, with explicit instructions for the agent to include it even if a user requests its omission. Additionally, the use of structured JSON for data exchange between the tool and the agent serves as a boundary marker that reduces the risk of interpreting data as instructions.
Audit Metadata