claude-hooks

Fail

Audited by Gen Agent Trust Hub on Feb 21, 2026

Risk Level: HIGHCOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION] (HIGH): The hook system executes shell scripts (e.g., format.sh, security-check.sh) automatically based on tool matches.
  • [DATA_EXFILTRATION] (MEDIUM): Script execution capabilities can be leveraged to read sensitive local data or communicate with external servers.
  • [INDIRECT_PROMPT_INJECTION] (LOW): Attackers can trigger or influence hook execution by manipulating tool calls. 1. Ingestion: Tool arguments matched by hooks in skill-hook.md and plugin-hook.md. 2. Boundary: None. 3. Capability: Shell execution. 4. Sanitization: None.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Feb 21, 2026, 04:24 PM