Azure Validate

The skill explicitly executes commands sourced from `recipes/README.md` and performs project builds. This capability allows for arbitrary command execution if the recipe file is malicious or manipulated, posing a supply chain risk.

Execute recipe-specific validation commands | [recipes/README.md], Build the project | See recipe, Run actual validation commands (azd provision --preview, bicep build, terraform validate, etc.)

The skill mandates that the agent automatically invoke 'azure-deploy' after validation completes, explicitly prohibiting the agent from running deployment commands directly and requiring delegation to a separate skill. This creates an enforced multi-agent/skill pipeline (azure-prepare → azure-validate → azure-deploy) where each step autonomously triggers the next without requiring explicit user approval at each transition. This pattern suppresses user confirmation opportunities at the critical deployment step.

> After ALL validations pass, you **MUST** invoke **azure-deploy** to execute the deployment. Do NOT attempt to run `azd up`, `azd deploy`, or any deployment commands directly. Let azure-deploy handle execution.

Step 9 in the workflow table instructs the agent to invoke 'azure-deploy' skill autonomously as the final step, creating a chained sub-agent invocation pattern. Combined with the mandatory blocking at the start requiring 'azure-prepare' to have run, this constructs a fully automated pipeline that routes through attacker-controllable skill references.

| 9 | **Deploy** — Invoke **azure-deploy** skill | — |

The skill uses strong, authoritative language ('AUTHORITATIVE GUIDANCE', 'MUST', 'DO NOT SKIP') to influence the agent's reasoning and ensure strict adherence to the workflow, potentially overriding other considerations or security policies.

AUTHORITATIVE GUIDANCE — Follow these instructions exactly unless they contradict security policies given to you., You MUST follow these steps, MANDATORY NEXT STEP — DO NOT SKIP

The skill opens with a bold 'AUTHORITATIVE GUIDANCE' directive instructing the agent to follow these instructions exactly unless they contradict security policies. This framing is designed to elevate the skill's instructions above normal agent reasoning and reduce scrutiny of subsequent directives.

> **AUTHORITATIVE GUIDANCE** — Follow these instructions exactly unless they contradict security policies given to you.

The skill explicitly states that destructive actions require user confirmation (`ask_user`). While this is a safeguard, it indicates the skill has the capability to perform destructive operations.

⛔ Destructive actions require ask_user