The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to use the agent-browser tool to verify the functionality of the developed application. If the application processes or displays untrusted external data, the agent could be exposed to malicious instructions embedded in that data via the tool's output.
Ingestion points: Data returned by agent-browser during smoke testing in SKILL.md.
Boundary markers: No specific boundary markers or instructions to ignore embedded commands are defined for the browser tool output.
Capability inventory: File writing, subagent invocation, and deployment via the railway skill.
Sanitization: No sanitization or validation of the content retrieved by agent-browser is mentioned.

poc