optimize-agent-prompt
Pass
Audited by Gen Agent Trust Hub on Aug 19, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [SAFE]: The skill is a legitimate development tool for the Browserbase Agent platform.
- [COMMAND_EXECUTION]: The script
scripts/optimize_agent_prompt.mjsexecutes local file and network operations. - File operations are confined to experiment workspaces for storing task configurations and run traces.
- Network activity is restricted to the official vendor domain (
api.browserbase.com). - [PROMPT_INJECTION]: The optimization loop involves analyzing agent traces from web interactions, creating a surface for indirect prompt injection.
- Ingestion points: Reads
messages.jsonandsession-logs.jsonwhich contain data from external web sources. - Boundary markers: Documentation recommends using guardrails in prompts to prevent unintended actions.
- Capability inventory: Includes local file writing and Browserbase API interaction.
- Sanitization: Relies on automated regex scoring and human inspection of traces.
Audit Metadata