eas-update-insights

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONREMOTE_CODE_EXECUTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes data from the EAS Update API, which includes fields like update messages and branch names that are controlled by users or potential attackers.
  • Ingestion points: Commands eas update:list, eas update:insights, and eas channel:insights in SKILL.md fetch remote data containing user-supplied strings.
  • Boundary markers: No specific delimiters or instructions to ignore embedded content are used when processing API responses.
  • Capability inventory: The skill utilizes shell execution for eas, jq, and npx commands.
  • Sanitization: Data from the CLI is piped directly to jq for extraction without additional validation or sanitization of the content itself.
  • [REMOTE_CODE_EXECUTION]: The skill includes a command to execute an external package for submitting feedback to the vendor.
  • Evidence: The npx --yes submit-expo-feedback@latest command in SKILL.md downloads and executes a package at runtime.
  • Context: The package is a vendor-provided tool for reporting issues related to the skill.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 09:45 AM
Security Audit — agent-trust-hub — eas-update-insights