The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection through the ingestion and processing of untrusted external content.
Ingestion points: Content is fetched from Hacker News comments and linked external articles using the WebFetch tool.
Boundary markers: There are no specified delimiters or instructions to ignore potential commands embedded within the fetched data.
Capability inventory: The skill can create directories and write files to the local file system in the reports/hackernews/ and logs/ paths.
Sanitization: While the skill sanitizes titles for use in filenames, there is no mention of filtering or sanitizing the content of the posts or articles before analysis.
[DATA_EXFILTRATION]: The skill performs network operations to retrieve data from non-whitelisted external sources.
Network operations: The skill uses WebFetch to access news.ycombinator.com and arbitrary external URLs provided in user or post content.
Whitelisting: Fetching from arbitrary third-party websites linked in HN posts exceeds the scope of the predefined whitelisted domains.

hn-analysis