The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted content from external sources to update the knowledge base, creating a surface for potential instruction injection.
Ingestion points: The skill pulls data from GitHub release bodies and pull request descriptions, as well as Linear project tracker details.
Boundary markers: There are no explicit delimiters or safety instructions provided to the agent to help it distinguish between the ingested data and its core instructions.
Capability inventory: The skill can execute shell commands (gh, mkdir), write to the local file system (05-knowledge/), and update external wiki pages through Notion and Confluence integrations.
Sanitization: The skill does not explicitly describe any validation or filtering of the external data before it is incorporated into the generated markdown files.
[COMMAND_EXECUTION]: The skill uses system commands and CLI tools to gather information and manage local files.
Evidence: Executes date for timestamps, mkdir for managing directories, and gh (GitHub CLI) for retrieving repository data. These operations are performed using standard tools and are aligned with the skill's primary purpose.

update-knowledge-base