The Agent Skills Directory

[DATA_EXFILTRATION]: The tool includes a --deliver webhook:<url> feature that allows the output of any command to be POSTed to an arbitrary external URL. While designed for data routing, this creates a potential vector for data exfiltration if the agent is directed to process sensitive information through the CLI.
[EXTERNAL_DOWNLOADS]: The skill instructions direct the installation of the hackernews-pp-cli binary using npx -y @mvanhorn/printing-press or go install from github.com/mvanhorn/printing-press-library. These resources originate from the identified developer's infrastructure.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection. It retrieves and processes unvalidated user-generated content from Hacker News (such as story titles and comment bodies).
Ingestion points: Data enters the agent's context through commands like items, stories, search, pulse, and hiring (SKILL.md).
Boundary markers: The tool uses --json and --agent flags to provide structured data, which helps delineate content but does not prevent the agent from interpreting instructions embedded within text fields.
Capability inventory: The skill has access to the Read Bash tool, allowing it to execute local CLI commands, perform network operations, and write to a local SQLite database.
Sanitization: There is no evidence of sanitization or filtering of the text content fetched from the Hacker News API before it is presented to the agent.

pp-hackernews