The Agent Skills Directory

[COMMAND_EXECUTION]: The skill documentation describes the use of the twg CLI tool to list object schemas, query asset types, and update settings for cross-schema relations in the Assets/CMDB environment. These operations are intended for managing operational data within the Atlassian infrastructure.
[EXTERNAL_DOWNLOADS]: The skill retrieves meeting transcripts, Jira work items, and JSM incident records through the twg tool as part of its data collection process for summaries and investigations.
[INDIRECT_PROMPT_INJECTION]: The skill processes content from external sources including incident comments, meeting transcripts, and Slack threads.
Ingestion points: twg jsm incident, twg jira workitem get, and meeting transcript retrievals.
Boundary markers: None explicitly defined in the prompt instructions.
Capability inventory: Limited to twg tool operations for querying, retrieval, and specific configuration updates (e.g., Assets schema settings).
Sanitization: No specific sanitization or escaping of external content is mentioned. However, the instructions emphasize classifying evidence by strength (e.g., 'hint', 'supported', or 'confirmed') and require establishing causal mechanisms before asserting facts, which serves as a significant guardrail against obeying instructions embedded in retrieved data.

twg-operational-health