peekaboo
Pass
Audited by Gen Agent Trust Hub on Mar 30, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: Installs the
peekabooCLI tool via a Homebrew tap (steipete/tap/peekaboo). This is an external binary dependency required for the skill's operation. - [COMMAND_EXECUTION]: The skill executes extensive system-level automation commands through the CLI, including launching/quitting applications, managing windows, and running local automation scripts via the
peekaboo runcommand. - [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8) because it extracts text and metadata from the screen to drive agent actions.
- Ingestion points: Screen content and UI element labels captured via
peekaboo seeandpeekaboo image(SKILL.md). - Boundary markers: Absent; no explicit instructions are provided to the agent to ignore or delimit instructions found within the UI text.
- Capability inventory: High-impact capabilities including simulated typing (
type), clicking (click), clipboard access (clipboard), and application management (app). - Sanitization: None; the skill does not appear to sanitize or validate text found on screen before interpreting it.
Audit Metadata