switchbot
Pass
Audited by Gen Agent Trust Hub on Oct 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill facilitates interaction with the official SwitchBot Cloud API (api.switch-bot.com), which is a well-known service for smart home device management.
- [SAFE]: The skill includes explicit instructions for the user to secure their local credentials file using restrictive permissions (chmod 600), which is a security best practice for managing API keys on a local filesystem.
- [INDIRECT_PROMPT_INJECTION]: The script
scripts/switchbot.pyaccepts device identifiers and command parameters from the agent's context and interpolates them into API requests. This creates an attack surface where maliciously crafted data in a processed document could attempt to trigger unintended smart home actions. - Ingestion points: Command-line arguments in
scripts/switchbot.pyoriginating from agent prompts. - Boundary markers: None explicitly provided in the skill instructions to separate data from commands.
- Capability inventory: Network operations via HTTPS POST/GET to the SwitchBot API.
- Sanitization: None; parameters are passed directly into the request URL and JSON payload.
Audit Metadata