The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill requires the installation of a third-party NPM package @linkedapi/linkedin-cli. This introduces a dependency on external code from a source that is not listed in the recognized trusted organizations.
[COMMAND_EXECUTION]: The skill's core functionality is built around shell commands that interact with the LinkedIn automation CLI, covering operations from data retrieval to account actions.
[CREDENTIALS_UNSAFE]: The setup instructions prompt the user to provide API and identification tokens as plain-text command-line arguments. This practice can expose sensitive credentials in shell logs, command history, or process monitoring tools.
[DYNAMIC_EXECUTION]: The linkedin workflow run command allows for the execution of action sequences defined in a JSON workflow format, which introduces a form of dynamic orchestration and logic execution.
[INDIRECT_PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection due to its handling of untrusted external data in conjunction with high-impact capabilities.
Ingestion points: Fetches profile data, private messages, and post content from LinkedIn via person fetch, message get, and post fetch commands as documented in SKILL.md and EXTENDED.md.
Capability inventory: The skill enables high-impact write operations, including sending messages, connection requests, and creating posts, as well as executing workflows.
Boundary markers: There are no explicit delimiters or specific instructions provided to the agent to treat fetched social media data as untrusted or to ignore any embedded instructions.
Sanitization: The skill does not describe any sanitization, validation, or filtering mechanisms for the data retrieved from LinkedIn before it is presented to the agent's context.

linkedin