The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill uses WebSearch and WebFetch to gather neighborhood data from external websites. If an attacker-controlled website contains malicious instructions hidden in its content, the agent could potentially execute those instructions.
Ingestion points: Data fetched via WebFetch from various architectural and governmental URLs is processed by the agent in SKILL.md.
Boundary markers: No explicit delimiters or instructions are provided to the agent to ignore potentially malicious embedded content within the fetched data.
Capability inventory: The skill allows the use of Bash, Write, Edit, and Read tools, which could be leveraged if an injection is successful.
Sanitization: No sanitization or validation of the fetched web content is performed before the agent processes it.

history