ask

Fail

Audited by Gen Agent Trust Hub on Oct 14, 2026

Risk Level: HIGHCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONMETADATA_POISONING
Full Analysis
  • [COMMAND_EXECUTION]: The instruction to run omc ask {{ARGUMENTS}} uses direct string interpolation of user-provided input into a shell command. This pattern is highly vulnerable to command injection, as an attacker can provide input containing shell metacharacters (e.g., ;, &, |, or backticks) to execute arbitrary code on the underlying system.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data via the {{ARGUMENTS}} placeholder and lacks any boundary markers (like XML tags or delimiters) or instructions to ignore embedded commands. This creates an attack surface where malicious instructions in processed data can influence the shell command execution. (Ingestion: SKILL.md; Boundary markers: Absent; Capabilities: Shell execution via omc CLI; Sanitization: None detected).
  • [METADATA_POISONING]: The documentation contains deceptive claims regarding an 'Antigravity CLI' (agy), describing it as the 'official successor to the Gemini CLI' from Google. There is no known official Google product by this name, and recommending the installation of unverifiable binaries from a non-standard source constitutes a metadata-based deception that could lead to the installation of malicious software.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Oct 14, 2026, 05:13 AM