agy-jobs
Pass
Audited by Gen Agent Trust Hub on Oct 6, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses a local Node.js companion script (
agy-companion.mjs) to perform all its operations, including job management and environment setup. - [INDIRECT_PROMPT_INJECTION]: The skill acts as an orchestrator that ingests output from background worker jobs. While this presents a potential attack surface for indirect prompt injection, the skill relies on standard tool-call patterns and provides guidance for handling results and errors.
- [PRIVILEGE_ESCALATION]: Documentation addresses sandbox limitations that might block access to sensitive local files (such as OAuth tokens) and instructs the agent to request explicit user permission for escalated access if required to bypass these restrictions.
- [SAFE]: The
setuputility is designed to modify global configuration files to establish permission rules. It follows best practices by using explicit user confirmation steps (AskUserQuestion) and providing a dry-run mode before applying any permanent changes to the system.
Audit Metadata