The Agent Skills Directory

[DATA_EXFILTRATION]: The script sets process.env.NODE_TLS_REJECT_UNAUTHORIZED = "0" and configures the database client with rejectUnauthorized: false. These settings globally disable SSL/TLS certificate validation for the process, exposing sensitive LinkedIn session cookies and database connection strings to potential Man-in-the-Middle (MITM) attacks.
[COMMAND_EXECUTION]: The skill utilizes child_process.execSync to execute shell commands such as clawdbot browser start to initialize the browser environment.
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface (Category 8) because it ingests and processes notification content from LinkedIn, which can be influenced by external actors.
Ingestion points: Notification text and links scraped from linkedin.com/notifications/ in scripts/linkedin-notifications.js.
Boundary markers: No delimiters or safety instructions are used to separate the untrusted scraped content from the agent's internal logic.
Capability inventory: The skill possesses database write permissions, network access via WebSockets and Fetch, and the ability to execute shell commands.
Sanitization: The code performs basic string cleaning and truncation but does not implement validation or escaping to prevent the execution of instructions embedded within notification text.

linkedin-notifications