sound-hooks

Pass

Audited by Gen Agent Trust Hub on Apr 2, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill directs the agent to insert shell commands into the 'hooks' section of the '~/.claude/settings.json' configuration file. These commands are configured to run the macOS 'afplay' utility in the background (e.g., 'afplay /System/Library/Sounds/Glass.aiff &') when specific agent events occur.- [COMMAND_EXECUTION]: The skill uses the 'jq' command-line tool to perform validation of the JSON structure within the settings file after updates are made.- [PROMPT_INJECTION]: The skill identifies an attack surface for indirect prompt injection because it allows user-supplied absolute file paths for custom audio files to be concatenated into shell command strings. * Ingestion points: User-provided absolute paths for custom .aiff, .mp3, or .wav files in the 'SKILL.md' instructions. * Boundary markers: None; the paths are interpolated directly into a command string. * Capability inventory: Capability to execute shell commands via 'hooks' in the settings file and direct execution of 'jq' for validation. * Sanitization: None; the skill instructions do not include steps to sanitize or escape shell metacharacters in the user-provided path string.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 2, 2026, 09:33 AM