The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the agent to install an external package @linkedapi/linkedin-cli from the npm registry.
[COMMAND_EXECUTION]: The skill makes extensive use of a CLI tool to perform network-based operations, including a specific command linkedin workflow run that executes logic defined in external JSON files or stdin.
[CREDENTIALS_UNSAFE]: The setup process requires the agent to handle and input sensitive 'Linked API' and 'Identification' tokens into the CLI tool, which may be exposed in command history or process logs.
[PROMPT_INJECTION]: The skill is vulnerable to Indirect Prompt Injection due to its data processing capabilities.
Ingestion points: Data is ingested from external LinkedIn profiles, messages, and posts via person fetch, message get, and post fetch commands.
Boundary markers: There are no instructions provided to use delimiters or protective markers to separate ingested content from system instructions.
Capability inventory: The skill possesses sensitive capabilities including message send, post create, and workflow run (which can execute arbitrary workflow steps).
Sanitization: There is no evidence of sanitization, filtering, or validation of the content retrieved from LinkedIn before it is presented to the agent.

linkedin-cli