The Agent Skills Directory

Remote Code Execution (CRITICAL): The skill explicitly instructs users to run curl -sSL https://canifi.com/skills/bitwarden/install.sh | bash and curl -sSL https://canifi.com/install.sh | bash. This piped remote execution pattern from an untrusted source allows for arbitrary code execution on the user's host machine.
Credentials Unsafe (HIGH): The setup process involves hardcoding or setting environment variables for highly sensitive secrets, including BW_SESSION, BW_CLIENTSECRET, and SERVICE_PASSWORD.
Data Exfiltration (HIGH): Because the skill has the capability to 'Get Secrets' and 'Send Files', a malicious script executed via the RCE vulnerability could programmatically dump and exfiltrate the user's entire Bitwarden vault.
Indirect Prompt Injection (LOW): The skill ingests untrusted data (passwords/usernames to be stored) through user prompts without visible boundary markers or sanitization, creating a surface for injection attacks, although the impact is secondary to the RCE. Ingestion points: Usage examples for creating/retrieving items. Boundary markers: Absent. Capability inventory: Bitwarden CLI access and file sharing. Sanitization: Absent.

bitwarden