managing-streamlit-apps
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFEDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [DYNAMIC_EXECUTION]: The skill describes the process of deploying and executing Python code in a managed sandbox environment via the
streamlit-apps-set-sourceandstreamlit-apps-starttools. This is the intended functional purpose of the skill and the associated vendor tools.\n- [INDIRECT_PROMPT_INJECTION]: The skill provides a mechanism for an agent to generate code and process data files (CSV, JSON, Parquet) based on user requests, which constitutes a potential attack surface.\n - Ingestion points: The
streamlit-apps-set-sourcetool accepts arbitrary Python source code and data assets (mapped in thefilesandassetsparameters) provided through user interactions.\n - Boundary markers: Instructions mandate the use of the
posthog_apps.query()interface and explicitly forbid standard imports (import posthog) to ensure compliance with sandbox security models.\n - Capability inventory: The skill utilizes tools for app creation, source code deployment, process control (start/stop), and deletion within the PostHog platform.\n
- Sanitization: The environment implements runtime restrictions, including a 30-second execution limit and a 256 MB memory cap per query, as part of its internal security posture.
Audit Metadata