The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests and processes content from repository files, creating a vulnerability surface for indirect prompt injection.\n
Ingestion points: The skill reads directory structures, SUMMARY.md navigation files, .gitbook.yaml configuration files, and various markdown pages from local environments or Git repositories.\n
Boundary markers: The instructions do not define specific delimiters or instructions for the agent to ignore embedded commands within the documentation content it processes.\n
Capability inventory: The agent has the capability to execute shell commands (gh, curl), write or rename files, and interact with the GitBook API.\n
Sanitization: There is no mention of sanitizing file content before it is processed or used in shell commands or logic.\n- [COMMAND_EXECUTION]: The skill provides instructions for the agent to execute specific CLI commands to manage the documentation workflow.\n
Evidence: Instructions include using the gh CLI to fetch commit statuses from GitHub and curl to interact with GitLab and GitBook APIs for retrieval of space IDs and page paths (e.g., in references/git-sync-previews.md and references/markdown.md). These commands are task-relevant and target well-known, trusted services.

write-docs