claude-code-hooks

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
  • [COMMAND_EXECUTION] (LOW): The skill's primary purpose is to help users configure shell-based automation hooks. It provides templates for executing commands like sh -c, git, npm test, and osascript. This is the intended functionality of the 'Claude Code hooks' feature.
  • [DATA_EXPOSURE] (LOW): The 'Session Context Injection' pattern in references/patterns.md suggests reading from ~/.claude/project-context.md. While this exposes local file content to the AI agent, it is presented as a mechanism for providing necessary project background to the LLM.
  • [INDIRECT_PROMPT_INJECTION] (LOW): The skill facilitates the ingestion of external data into the agent's context, creating a potential surface for indirect injection.
  • Ingestion points: ~/.claude/project-context.md (via the Session Start hook).
  • Boundary markers: None shown; content is read directly into the session.
  • Capability inventory: The skill manages hooks that can execute arbitrary shell commands, write files, and trigger notifications.
  • Sanitization: None; the patterns rely on the user to ensure injected context is safe.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:15 PM