The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the olares-cli binary for all operations, including profile management, application lifecycle monitoring, and storage interaction.
Evidence includes commands like olares-cli profile list, olares-cli profile login, and olares-cli settings me version used to interact with the Olares platform.
[EXTERNAL_DOWNLOADS]: The skill provides instructions for installing the vendor's command-line interface via the NPM registry.
Evidence: npm install -g @olares/cli@latest found in references/olares-suite-map.md. This targets the official package for the 'beclab' vendor.
[INDIRECT_PROMPT_INJECTION]: The skill establishes a protocol for parsing machine-readable JSON output from the CLI, which represents an ingestion surface for external data.
The instructions provide specific guidance on branching based on error codes (auth_no_profile, auth_not_logged_in, etc.) to ensure predictable agent behavior when processing tool output.
[CREDENTIALS_UNSAFE]: The skill explicitly enforces secure credential management practices.
Evidence: Explicit warnings in SKILL.md to "Never place a password in command arguments" and "Never print access or refresh tokens." It also documents the transition from deprecated plaintext storage (tokens.json) to secure OS-level backends like macOS Keychain and Windows DPAPI.

olares-shared