The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: Recommends the installation of the @linkedapi/linkedin-cli package from the public NPM registry and communicates with the app.linkedapi.io service to perform LinkedIn automation.
[COMMAND_EXECUTION]: The skill's primary functionality is built around executing shell commands using the linkedin CLI tool for profile fetching, messaging, and account management.
[CREDENTIALS_UNSAFE]: The setup process requires users to pass sensitive API and identification tokens as plain-text arguments to the linkedin setup command. This can result in credentials being stored in the user's shell history (e.g., ~/.bash_history), which is a security risk.
[PROMPT_INJECTION]: The skill processes untrusted data from LinkedIn, including profiles, posts, and messages. This creates an attack surface for indirect prompt injection:
Ingestion points: Content is retrieved via commands like person fetch, message get, and post fetch in SKILL.md.
Boundary markers: No delimiters or instructions are specified to prevent the agent from obeying instructions embedded in the retrieved data.
Capability inventory: The skill has extensive command execution capabilities through the linkedin tool.
Sanitization: There is no mention of sanitizing or escaping the retrieved external content before it is processed by the agent.

linkedin-cli