hook-development

Pass

Audited by Gen Agent Trust Hub on Feb 25, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No malicious patterns, hidden commands, or unauthorized data exfiltration were detected. The skill is designed to improve developer productivity and security during hook implementation.
  • [COMMAND_EXECUTION]: Utility scripts such as test-hook.sh, hook-linter.sh, and validate-hook-schema.sh execute shell commands to facilitate testing and static analysis of hooks. This functionality is intended for local developer use.
  • [PROMPT_INJECTION]: The skill describes an indirect prompt injection surface (Category 8) where untrusted data is processed at runtime.
  • Ingestion points: Untrusted data enters via placeholders such as $TOOL_INPUT, $TOOL_RESULT, and $USER_PROMPT in prompt-based hooks.
  • Boundary markers: Absent in provided examples; values are interpolated directly into prompt strings.
  • Capability inventory: Hooks possess the capability to execute shell scripts (command type) or block tool execution (prompt type).
  • Sanitization: Examples demonstrate field extraction with jq but do not include explicit escaping or sanitization of content before prompt interpolation.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 25, 2026, 04:46 PM