Skills
Navigate
skills/jackwener/opencli/opencli-usage/Gen Agent Trust Hub

opencli-usage

Fail

Audited by Gen Agent Trust Hub on Oct 10, 2026

Risk Level: HIGHREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTIONPERSISTENCE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill documents the opencli plugin install github:user/repo command, which allows the installation of third-party plugins from arbitrary GitHub repositories.
  • [REMOTE_CODE_EXECUTION]: The documented plugin installation mechanism enables the execution of code from untrusted external sources.
  • [INDIRECT_PROMPT_INJECTION]:
  • Ingestion points: The skill utilizes opencli <site> <command> and opencli browser subcommands to ingest data from external websites into the agent's context.
  • Boundary markers: No explicit boundary markers or "ignore embedded instructions" warnings are documented for processing website data.
  • Capability inventory: The skill has the capability to execute bash commands (opencli:*) and modify local file systems and shell configurations.
  • Sanitization: No sanitization, escaping, or filtering of external website content is described.
  • [PERSISTENCE]: The opencli completion bash command facilitates the modification of shell profiles (e.g., ~/.bashrc) to add persistent autocomplete scripts.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Oct 10, 2026, 12:36 PM