Skills
Navigate
skills/gupsammy/meta-ads-cli/meta-ads-intel/Gen Agent Trust Hub

meta-ads-intel

Fail

Audited by Gen Agent Trust Hub on Sep 29, 2026

Risk Level: HIGHPERSISTENCEPRIVILEGE_ESCALATIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [PERSISTENCE]: The skill provides instructions to set up persistent automated tasks using macOS launchd plists or Linux crontab entries to run the analysis weekly.
  • [PRIVILEGE_ESCALATION]: The onboarding documentation suggests that users execute sudo npm i -g meta-ads if they encounter permission issues during installation.
  • [EXTERNAL_DOWNLOADS]: The skill downloads the meta-ads package from the NPM registry as a required dependency.
  • [COMMAND_EXECUTION]: The skill runs shell commands using the meta-ads CLI and passes user-supplied $ARGUMENTS directly into the execution string.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from Meta API responses and deep scrapes of user-provided website URLs using a subagent. It lacks boundary markers to separate external data from instructions and does not perform sanitization of the ingested content, while possessing shell execution and file-write capabilities.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Sep 29, 2026, 12:50 AM