Microsoft Foundry Skill

The skill explicitly uses `azd env get-values` and Azure CLI for resource creation, and implicitly performs container operations (build, push to ACR). This grants the agent broad command execution capabilities in the underlying system or cloud environment, which could be exploited for arbitrary code execution or system compromise.

run `azd env get-values` and use it to seed `agent-metadata.yaml`... Creating Azure AI Services multi-service resource (Foundry resource) using Azure CLI... Containerize, build, push to ACR

The skill reads and resolves sensitive configuration values like project endpoints, ACR registries, and Azure subscription IDs from local files (`agent-metadata.yaml`) and system commands (`azd env get-values`). It also accesses traces, logs, and RBAC information, posing a significant risk of sensitive data exposure or exfiltration if the agent is compromised.

`agent-metadata.yaml` is the required source of truth for environment-specific project settings, agent names, registry details... run `azd env get-values` and use it to seed `agent-metadata.yaml`... Query traces, analyze latency/failures... Managing RBAC permissions

The skill explicitly uses the `task` or `runSubagent` tool to delegate sub-tasks. This capability can be abused to launch attacker-controlled sub-agents, potentially leading to compositional attacks, privilege escalation, or uncontrolled execution if not properly sandboxed or if sub-agent prompts are manipulated.

Use the `task` or `runSubagent` tool to delegate long-running or independent sub-tasks