The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill establishes a workflow for capturing and processing untrusted user feedback which is subsequently used to drive modifications to the Power BI semantic model and report definitions.
Ingestion points: External user feedback is ingested via the structured templates defined in references/feedback-intake-template.md and references/classification.md.
Boundary markers: Absent. The skill provides no instructions for the agent to use delimiters or ignore instructions embedded within the user-provided feedback string.
Capability inventory: Feedback items are routed to downstream skills (e.g., power-bi-semantic-model, power-bi-dax-development) which possess capabilities to modify TMDL and JSON files in the repository.
Sanitization: Absent. There are no guidelines for validating or escaping user-provided feedback before it is passed to implementation agents.
[COMMAND_EXECUTION]: The skill provides several executable shell command snippets intended for version control and environment management.
Evidence: references/ab-variant-testing.md contains complex Git workflows for creating, pushing, and merging feature branches.
Evidence: references/git-pbip-diff-guide.md references the use of powerbi-report-author validate and local validation scripts for structural integrity checks.

power-bi-feedback-iteration