The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses a dedicated command-line tool, reader-cli, to perform its operations. This is the primary mechanism for the skill's functionality.
Evidence includes multiple bash commands such as reader-cli whoami, reader-cli add <url>, and reader-cli list.
[CREDENTIALS_SAFE]: The skill includes explicit instructions for handling authentication securely.
It advises checking login state with reader-cli whoami --json before accessing data.
It prioritizes interactive terminal logins and explicitly warns: "Never print the full API key in conversation or command output."
[INDIRECT_PROMPT_INJECTION]: The skill processes external data (URLs and bookmark content via reader-cli get <bookmark-id> --markdown).
Ingestion points: Bookmark content and URLs fetched via reader-cli.
Boundary markers: None explicitly mentioned in prompt construction, but the skill uses structured JSON output for command results.
Capability inventory: File system access and network operations are delegated to the reader-cli binary rather than direct shell scripts.
Sanitization: Not explicitly mentioned, relying on the CLI tool and LLM guardrails.

slax-reader