The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides specific shell commands using gh (GitHub CLI), curl, and jq for the agent to execute. These commands are used to retrieve commit statuses, resolve space IDs, and fetch page paths necessary for providing site preview links and maintaining cross-space documentation references.
[EXTERNAL_DOWNLOADS]: The skill instructs the agent to interact with well-known and trusted APIs, including api.gitbook.com, GitHub, and GitLab. These interactions are scoped to retrieving documentation metadata and deployment status.
[INDIRECT_PROMPT_INJECTION]: The skill's primary function involves reading and writing documentation files (Markdown, YAML), which are external data sources. This establishes an ingestion point for untrusted content. The skill provides specific behavioral instructions, such as the 'two-link rule' and SUMMARY.md verification, to guide the agent in processing these files correctly.
[CREDENTIALS_UNSAFE]: The documentation references the use of GITBOOK_TOKEN and GITLAB_TOKEN for API authentication. It correctly advises the agent to use these as environment variables or within authorization headers rather than hardcoding values into scripts or documents.

write-docs