ultra-research

Pass

Audited by Gen Agent Trust Hub on Aug 14, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes local Node.js scripts (auth.js, grep-api.js) to check authentication status and interact with the GREP research API. These scripts are resolved relative to the skill directory, targeting the vendor's own tooling.
  • [PERSISTENCE_MECHANISMS]: Orchestrates recurring tasks via the /loop command to poll for job results every 5 minutes. This allows the research to continue across agent sessions. The instructions include mandatory steps to delete the scheduled task using CronDelete once the job completes or fails.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data returned from an external API research job, which could potentially contain adversarial instructions.
  • Ingestion points: The output of node grep-api.js result <job_id> processed in SKILL.md and the recurring loop.
  • Boundary markers: Absent; the instructions do not specify the use of delimiters or 'ignore' instructions when presenting the external report.
  • Capability inventory: The skill possesses the ability to manage cron jobs (CronDelete, CronList) and execute shell scripts.
  • Sanitization: No explicit sanitization or filtering of the API output is described before it is presented to the user.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 14, 2026, 09:40 PM
Security Audit — agent-trust-hub — ultra-research