k6-manage

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill explicitly tells the agent to prompt the user for an API token/stack and shows examples embedding that token verbatim into curl/CLI headers and commands (Authorization: Bearer $K6_TOKEN, --token "$TOKEN"), which requires the LLM to handle/output secrets directly.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.30). Required runtime workflow reads outsider-authored free text via audit/result fields in Grafana Cloud k6 “Cloud Insights” when it calls /api/plugins/k6-app/resources/insights/insights/api/v1/testrun/$RUN_ID/executions, /audits, and /audits/results and then joins fields like explanation and actions into the agent context.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly instructs the agent to fetch the OpenAPI spec at "/api/plugins/k6-app/resources/cloud/cloud/v6/openapi" (gcx api /.../openapi) to discover operations at runtime and to fetch test scripts from "/api/plugins/k6-app/resources/cloud/cloud/v6/load_tests/$ID/script" (and via direct curl to https://api.k6.io/cloud/v6/auth in examples); these runtime fetches supply schemas and executable script bytes that control agent behavior and can be executed locally (k6 run), so they are runtime external dependencies that affect prompts/execution.