The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion of content from external websites via web search and fetching tools. This allows for potential indirect prompt injection, where an attacker could embed malicious instructions in a webpage (e.g., a fake funding announcement) to influence the agent's behavior.
Ingestion points: External content is fetched from federal agency sites (NSF, NIH, DOE, etc.), foundations, and grant aggregators using web tools as defined in SKILL.md.
Boundary markers: There are no specific instructions to use delimiters or ignore instructions found within the external content during the summarization and report generation process.
Capability inventory: The agent has access to web_search and web_fetch tools, as well as the ability to write the final report to a local file in /mnt/user-data/outputs/.
Sanitization: The skill lacks requirements for sanitizing or validating the text retrieved from external sources before it is processed and formatted.

funding-search