The Agent Skills Directory

[SAFE]: No security issues detected. The skill's behavior is consistent with its stated purpose of providing semi-automated fact-checking and media literacy education.
[INDIRECT_PROMPT_INJECTION]: The skill possesses a data ingestion surface as it fetches content from user-provided URLs and analyzes untrusted text/images. While this is a known vector for indirect prompt injection, it is essential for the skill's primary function (fact-checking) and is handled through a structured analysis pipeline rather than direct execution.
Ingestion points: External URLs (via web_fetch), user-supplied text, and extracted text from images.
Boundary markers: The skill does not explicitly define delimiters for fetched web content, which is a common limitation for analysis-focused tools.
Capability inventory: Limited to web_search, web_fetch, and image processing. The skill does not have the capability to write files or execute system commands.
Sanitization: The skill focuses on analytical decomposition of claims rather than raw interpolation, reducing the risk of accidental instruction following.
[EXTERNAL_DOWNLOADS]: The skill methodology references fetching data from well-known and trusted services such as Google Search, Wikipedia, and established international fact-checking organizations (Snopes, Reuters, WHO, etc.). These references are used solely for information gathering and do not involve the execution of remote scripts or binary payloads.

fact-check