The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill fetches canonical guidance and product configuration from Elastic's official GitHub repositories and documentation website.
Resources fetched include ChangelogEntry.cs and products.yml from the elastic/docs-builder repository.
Standards are also retrieved from the official elastic.co documentation site.
[INDIRECT_PROMPT_INJECTION]: The skill reads and processes untrusted changelog YAML files from the workspace, creating a potential surface for embedded instructions to influence the agent's behavior.
Ingestion points: The skill glob-searches for and reads *.yaml and *.yml files from a user-provided path or directory (Step 2).
Boundary markers: The instructions do not specify the use of delimiters or 'ignore embedded instructions' warnings when passing changelog content to the agent for review.
Capability inventory: The skill has access to the WebFetch tool for external data retrieval and the Read tool for accessing local files.
Sanitization: There is no evidence of sanitization or filtering applied to the text fields within the changelog files before they are analyzed by the AI.

docs-review-changelog