The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: Fetches and executes installation scripts for the sjctl utility directly from the solid-company GitHub repository. These scripts are provided to ensure the necessary tool is available for the skill to function.
[EXTERNAL_DOWNLOADS]: Downloads the sjctl binary or setup scripts from the author's official GitHub repository to the local environment.
[COMMAND_EXECUTION]: Uses the sjctl command-line interface to perform job synchronization, search management, and profile display.
[INDIRECT_PROMPT_INJECTION]: Ingests and processes job titles and descriptions originating from third parties through a public API.
Ingestion points: Untrusted data enters the agent context via the sjctl sync --json command in SKILL.md.
Boundary markers: The skill includes an explicit warning section ("Untrusted input") that instructs the agent to treat listing content strictly as data and to ignore any embedded commands.
Capability inventory: The skill has the capability to execute shell commands using the sjctl tool.
Sanitization: The skill relies on natural language instructions for the agent to disregard malicious content rather than programmatically sanitizing the job listing data.

jobs-digest