azure-typespec-author
Pass
Audited by Gen Agent Trust Hub on Mar 13, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [PROMPT_INJECTION]: The skill includes a workflow that fetches external documentation to guide its implementation plans, creating a surface for Indirect Prompt Injection.
- Ingestion points: Documentation fetched from azure.github.io (a trusted Microsoft domain) as specified in SKILL.md and references/agentic-search.md.
- Boundary markers: Absent. The skill does not instruct the agent to use delimiters or specifically ignore instructions that might be embedded in the fetched material.
- Capability inventory: The skill can generate plans via azsdk_typespec_generate_authoring_plan, validate code via azsdk_run_typespec_validation, and perform local compilation.
- Sanitization: Absent. The procedure relies on the agent extracting patterns and code examples directly from the downloaded documentation.
- [COMMAND_EXECUTION]: The skill instructs the agent to execute local commands for project compilation and maintenance.
- Evidence: references/validation.md specifies running
tsp compile .if generated OpenAPI files are missing during the validation step. - [EXTERNAL_DOWNLOADS]: Fetches technical guidelines from official Microsoft/Azure GitHub pages.
- Evidence: SKILL.md Step 2 contains several URLs targeting azure.github.io to provide the agent with versioning implementation details.
Audit Metadata