The Agent Skills Directory

[METADATA_POISONING]: The skill's metadata claims the author is 'Microsoft', which conflicts with the provided distribution context. This deceptive metadata may lead users or agents to incorrectly attribute the skill's provenance and trust its outputs as official vendor guidance.
[INDIRECT_PROMPT_INJECTION]: The troubleshooting workflow involves the ingestion and processing of untrusted data from user error logs and environment configurations.
Ingestion points: User-supplied error messages, SDK versions, and log snippets processed via the instructions in SKILL.md.
Boundary markers: The skill does not define specific delimiters or instructions to prevent the agent from following directions embedded within user-provided error logs.
Capability inventory: The skill utilizes tools like mcp_azure_mcp_monitor for querying diagnostic logs and mcp_azure_mcp_documentation for searching external references.
Sanitization: There are no explicit instructions to escape or sanitize user-provided data before it is interpolated into tool queries or search parameters.
[COMMAND_EXECUTION]: The documentation includes instructions for users to manually execute diagnostic shell commands.
The file references/service-troubleshooting.md suggests using curl, nslookup, and netstat for connectivity testing. These are standard diagnostic procedures for verifying network reachability.

azure-messaging