audit-paper-book
Pass
Audited by Gen Agent Trust Hub on May 8, 2026
Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes
launchctlcommands to restart a local macOS Launch Agent (com.user.atlas). This action is intended to refresh the user's local research atlas after updates are applied, aligning with the skill's primary purpose. No elevation of privileges is requested. - [DATA_EXFILTRATION]: The skill uses
curlto interact withlocalhost:8770. These requests are used to verify that the local service is correctly serving updated research assets (references and figures). No data is transmitted to external or untrusted network domains. - [PROMPT_INJECTION]: The skill identifies a potential surface for indirect prompt injection as it processes content from LaTeX files (
.texand.bib) using regex and grep. However, because the skill's operations are limited to reporting and specific file copies without executing the file content, the risk is negligible. No explicit boundary markers are defined for the ingested content.
Audit Metadata