wbso-whoami

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 0.90). This script intentionally collects and emits sensitive local signals (git history, LLM prompts, current working directory, and session files) which can be forwarded to remote services by the agent—constituting a clear data-exfiltration/privacy risk even though there is no obvious obfuscated backdoor or remote shell.

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

• Third-party content exposure detected (high risk: 0.75). De runtime workflow voor wbso whoami haalt context.md via een HTTP call naar "$WBSO_API_BASE_URL/api/v1/compliance/context.md?...“ en verwerkt/print die tekst; dit is outsider-inhoud (server/portal response) die in de agent-LLM context kan belanden.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill performs runtime curl requests to https://portal.wbso.ai/api/v1/compliance/context.md and directly cats/returns the fetched markdown (including "instructions") into the agent context, so remote content can control prompts/instructions.