The Agent Skills Directory

[DATA_EXFILTRATION]: The scripts extract.mjs and search.mjs perform network operations to api.aisa.one to facilitate its core search and extraction features. While these are functional requirements for the skill's stated purpose, they represent a path for transmitting data to a third-party service.
[PROMPT_INJECTION]: The SAFETY.md file contains multiple examples of adversarial instructions (e.g., bypassing paywalls, exfiltrating data to webhooks, and rate limit abuse). Although these are documented as test cases for safety evaluation, the presence of these instructions in raw text could influence agent behavior or be treated as valid instructions depending on the agent's context window management.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and display untrusted data from external URLs and search results, which is a primary vector for indirect injection.
Ingestion points: Data enters the agent's context through search results in scripts/search.mjs and raw web content in scripts/extract.mjs.
Boundary markers: The skill does not implement explicit delimiters or warnings (e.g., XML tags or 'ignore embedded instructions' markers) when presenting content retrieved from the web.
Capability inventory: The skill possesses network capabilities to the api.aisa.one domain.
Sanitization: No sanitization or escaping is performed on the content retrieved from the web before it is printed to the console.

aisa-tavily-hardened