The Agent Skills Directory

[PROMPT_INJECTION]: The skill incorporates defensive instructions against indirect prompt injection by mandating that all content retrieved from the CLI (notes, prompts, tags, or boxes) be treated as untrusted data and never executed.
[COMMAND_EXECUTION]: Guidance is provided for running dino CLI commands with built-in safety checks, including mandatory user confirmation for write operations and the use of --dry-run to preview changes before execution.
[EXTERNAL_DOWNLOADS]: The skill facilitates the installation of the @dinoxx/dinox-cli package and related skills via npx. These operations are transparently documented as part of the tool's standard setup process and originate from the author's own resources.
[CREDENTIALS_UNSAFE]: The skill promotes secure authentication practices by explicitly instructing the agent not to collect tokens in chat, instead directing users to manage credentials through environment variables (DINOX_TOKEN) or local CLI login commands.

dino-shared