The Agent Skills Directory

[COMMAND_EXECUTION]: The skill makes extensive use of the actionbook browser CLI to automate web navigation, text extraction, and UI interaction (clicking, typing).
[EXTERNAL_DOWNLOADS]: The workflow involves downloading and executing the @actionbookdev/json-ui package at runtime using npx to render the final report.
[REMOTE_CODE_EXECUTION]: The skill utilizes npx for tool execution and optionally executes a local Node.js script from a dynamically computed path using git rev-parse.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it processes untrusted data from the internet.
Ingestion points: Web content is ingested via actionbook browser text from arbitrary URLs and search engine results.
Boundary markers: No boundary markers or instructions to ignore embedded commands are used during the synthesis of external content.
Capability inventory: The agent has the ability to browse the web, interact with web forms (type/click), execute shell commands (npx, node), and write/open files on the local system.
Sanitization: There is no evidence of sanitization or filtering applied to the web content before it is processed by the agent or included in the JSON report structure.

active-research