The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill fetches content from arbitrary, user-provided URLs and search results which are then ingested into the agent context. This creates a surface where malicious instructions embedded in web pages could attempt to influence the agent's behavior.
Ingestion points: Content is fetched from external sources via scripts/webtools.py using endpoints such as scrape, search, and crawl.
Boundary markers: The skill returns data in structured formats (JSON or Markdown), but it does not provide explicit instructions to the agent to treat field content as untrusted or to ignore embedded commands within the fetched data.
Capability inventory: The skill possesses the capability to read content from any public URL and perform search queries via the ZooData API.
Sanitization: No sanitization or instruction-filtering is performed on the scraped content by the skill script before it is returned to the agent context.
[DATA_EXFILTRATION]: The skill script scripts/webtools.py performs network operations targeting api.zoodata.ai to facilitate web scraping and search services. It also accesses a local configuration file at ~/.zoodata/config.json to retrieve necessary API credentials. These operations are consistent with the skill's documented purpose as an interface for the ZooData service provided by the vendor SerendipityOneInc.

web-extract