Security audit
Security checks for vulnerabilities and agentic risk
The skill’s purpose is simple and mostly coherent, but it ships code that uses shell-based curl despite documenting only safer web fetching.
Review before installing because the skill includes a Node script that invokes a shell command to fetch the page, while its instructions describe using web_fetch. The content source is public and fixed to wol.jw.org/es, but a safer version should remove shell execution or disclose and tightly scope it.
The declared behavior is limited to retrieving JW daily text via web_fetch, but the analysis reports use of child_process.execSync and curl instead. This mismatch is dangerous because shell-based network access bypasses declared tool restrictions, reduces auditability, and can open the door to arbitrary command execution or broader outbound access than intended.
The skill declares only a benign web-fetching purpose, but the analysis indicates shell capability without any explicit tool scoping or permissions declaration. Hidden execution capability increases risk because it can enable unintended command execution paths that are not visible to users or policy controls.
The instructions explicitly require always using the Spanish version of the site and prohibit translation. This imposes a fixed language/locale behavior without indicating that the user can choose another language, which matches the language/locale policy violation category.
The skill invokes a shell command via child_process to fetch remote content, which gives it broader execution capability than its stated purpose of simply retrieving a webpage. Even though the URL is locally constructed from the current date and not directly user-controlled here, using execSync with shell execution unnecessarily expands the attack surface and can enable abuse if the code is later modified or the environment is manipulated.
The comment explicitly states the URL is always in Spanish, and the code forces the /es/ path and es-ES locale formatting. This is a natural-language policy concern because it imposes a specific language without user opt-in or any documented justification that the skill is region-specific.
The implementation does not match the declared design: it uses curl through shell execution instead of the safer, expected web_fetch capability. This discrepancy matters because it bypasses intended platform controls, auditing, and restriction boundaries that may apply to approved fetch mechanisms, making the skill more dangerous than its description suggests.
Detected: suspicious.dangerous_exec
