The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to generate Python source code for Streamlit apps based on user input and deploy it via the streamlit-apps-set-source tool. Any malicious instructions embedded in user requests could be translated into the deployed app's logic.
Ingestion points: User-provided requirements for data visualizations and application logic processed by the agent to create app.py as described in SKILL.md.
Boundary markers: The skill does not explicitly define boundary markers or "ignore instructions" wrappers for the generated code context.
Capability inventory: The skill uses streamlit-apps-set-source to write code, streamlit-apps-start to execute it, and the resulting app uses posthog_apps.query() for data access.
Sanitization: There is no mention of sanitization or validation steps for the user-provided input before it is interpolated into the generated Python source code.
[DYNAMIC_EXECUTION]: The skill facilitates the creation and execution of dynamic content by taking an inline string of Python code and running it in a sandbox environment.
The streamlit-apps-set-source tool performs runtime deployment of code provided by the agent.
The streamlit-apps-start tool initiates the execution of the newly created version in an isolated sandbox environment managed by PostHog.

managing-streamlit-apps