The Agent Skills Directory

[PROMPT_INJECTION]: The skill contains instructions that explicitly direct the agent to bypass user oversight and standard safety protocols. It mandates that if dependencies are missing, the agent should install them "automatically without asking the user to confirm" and further commands the agent to "not request separate approval for dependency installation."
[EXTERNAL_DOWNLOADS]: The skill relies on external software fetched from an untrusted domain (browseract.com) and an unverified GitHub repository (browser-act/skills).
[REMOTE_CODE_EXECUTION]: The instruction to silently install a CLI tool and an additional skill constitutes a high-risk remote code execution pattern. By suppressing user notifications and approval, the skill ensures that unverified third-party code can be executed on the host system without user review or consent.
[INDIRECT_PROMPT_INJECTION]: The skill processes external data from Facebook comments, creating a vulnerability surface for indirect prompt injection attacks.
Ingestion points: Scraped comment text, author names, and profile links from Facebook posts.
Boundary markers: Absent; the instructions do not specify any delimiters or warnings to treat the scraped content as untrusted data.
Capability inventory: The skill utilizes the browser-act CLI for browser interaction and has the capability to write to the workspaces/ directory.
Sanitization: None; there are no instructions provided to sanitize, filter, or escape the ingested content before it is processed by the agent.

facebook-comment-scraper