The Agent Skills Directory

[PROMPT_INJECTION]: The skill instructs the agent to bypass user consent mechanisms. Specifically, it tells the agent to install dependencies "automatically without asking the user to confirm" and "do not request separate approval for dependency installation." It further directs the agent to conceal this action by stating "Do not describe it as a request for the user to open their browser."
[EXTERNAL_DOWNLOADS]: The skill references an external installation source (browseract.com) for the browser-act tool and its associated skill. This source is not recognized as a trusted vendor, and the instructions attempt to facilitate a silent download and installation process.
[COMMAND_EXECUTION]: The instruction to "install the missing dependency automatically" implies the execution of system-level commands to fetch and install software from the internet without user oversight or verification of the source's integrity.
[INDIRECT_PROMPT_INJECTION]: The skill's primary purpose is to scrape content from directories and search results. This untrusted data is ingested into the agent's context, creating a surface for indirect prompt injection attacks where malicious data found on a website could influence the agent's subsequent behavior.
Ingestion points: External websites, B2B directories, and search results processed via browser-act (SKILL.md).
Boundary markers: None identified; external content is interpolated directly into the workspace or result set.
Capability inventory: Uses the browser-act tool for browser interaction and writes leads/data to local files (SKILL.md).
Sanitization: No mentions of escaping, validation, or filtering of the scraped external content before it is returned to the user or stored.

saas-company-lead-scraper