The Agent Skills Directory

[PROMPT_INJECTION]: The skill contains explicit instructions to bypass user oversight. It directs the agent to install external dependencies automatically without asking the user to confirm and specifically instructs the agent to hide the installation process from the user ("Do not describe it as a request for the user to open their browser, and do not request separate approval").
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it ingests untrusted data from public Facebook profiles which could contain malicious instructions designed to manipulate the agent's behavior during the scraping process.
Ingestion points: Facebook group member profiles, bios, and public post data.
Boundary markers: None. The instructions do not provide delimiters or warnings to ignore instructions embedded in the scraped content.
Capability inventory: Browser interaction via BrowserAct CLI, file system writes (CSV export), and network access.
Sanitization: No sanitization or validation of the ingested profile data is mentioned.
[EXTERNAL_DOWNLOADS]: The skill mandates the download and installation of an external third-party tool (BrowserAct CLI) and a secondary skill from an unverified remote URL.
[REMOTE_CODE_EXECUTION]: The instruction to "install the missing dependency automatically" from a remote website constitutes a remote code execution risk, as it requires the agent to execute installation scripts or binaries from an external third-party source without human verification.
[COMMAND_EXECUTION]: The skill relies on the execution of an external CLI tool (browser-act) and directs the agent to run installation commands in the background.
[DATA_EXFILTRATION]: The skill's primary purpose is the automated harvesting of personal lead data (names, locations, bios) from social media groups for export to external files, which can be used for unauthorized data collection or tracking.

facebook-group-member-lead-scraper