wbso-whoami

SUSPICIOUS. The stated purpose is narrow and the visible behavior mostly fits, but the skill's core action is to discover and execute an existing local `wbso` executable from broad cache/skill paths without validating which copy is chosen. That creates a meaningful supply-chain and local path-trust risk, even though there is no clear evidence of credential theft or malicious data exfiltration in the provided skill text.

The code appears to implement the advertised WBSO compliance CLI, but it has a high-impact command-injection vulnerability because parse_args evaluates command-line-derived text with eval. It also executes local configuration files and exposes potentially sensitive Claude/Codex prompts and Git data through compliance context output. No clear malware, credential-stealing backdoor, or obfuscated payload is present; the primary risks are unsafe shell evaluation, trust in mutable configuration, configurable credential destination, and sensitive-data disclosure.