deploy

Runs verified deploys with healthchecks, event logging, and automatic rollback on failure.

Updated Jun 28, 2026
One-click install
npx skills add https://github.com/sicambria/talkteach-asr --skill deploy-sicambria
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: deploy
Source: https://github.com/sicambria/talkteach-asr/tree/main/.claude/skills/deploy
Command: npx skills add https://github.com/sicambria/talkteach-asr --skill deploy-sicambria

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Ad-hoc deploys often skip the two checks that catch most bad releases: confirming the codebase passes verification before shipping, and confirming the deployed service is actually healthy afterward. This Skill makes both steps mandatory and automatic, and rolls back when a healthcheck fails. ## Core Features & Use Cases - Fail-closed deploy pipeline: Runs the full verify contract first and aborts before touching the deploy command if anything is red. - Healthcheck with automatic rollback: Validates the deployed service via an HTTP endpoint or command, and runs a configured rollback command if the check fails. - Encrypted backups and audit logging: Optionally runs and AES-256-GCM-encrypts a pre-deploy backup, and logs every step to an event spine for later review. - Use Case: A CI/CD job or developer runs a single deploy entrypoint that verifies the tree, ships to production, healthchecks the service, and automatically rolls back a broken release instead of leaving it live. ## Quick Start Ask the assistant to deploy the current tree using the configured deploy command and healthcheck.

Frequently Asked Questions about deploy

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I set up an automated deploy with healthcheck and rollback?▼

Configure config.deploy in .harness/config.json with a deploy cmd, a healthcheckUrl or healthcheckCmd, and a rollbackCmd. Running the deploy then verifies the tree, deploys, healthchecks, and rolls back automatically if the healthcheck fails.

What happens if the deploy command is not configured?▼

The runner abstains: it logs an abstain event and exits non-zero with a clear message. It never reports success when there is nothing configured to run, so you must fill config.deploy.cmd to enable it.

Does the deploy run tests before shipping to production?▼

Yes. It always runs the full verify contract first via verify.mjs --phase full. If verification is red, the deploy aborts before the deploy command is ever executed.

How are pre-deploy backups encrypted and restored?▼

If backupCmd is set and KZ_DEPLOY_BACKUP_KEY is present, the newest file in .harness/state/backups/ is encrypted with AES-256-GCM and the plaintext deleted. Restore manually with a Node script using the SHA-256 of the key, since no automated restore ships.

Why did my deploy report failure even though rollback succeeded?▼

A failed healthcheck always marks the deploy as failed, regardless of whether the rollback command succeeded. Rollback only restores the previous state; it does not change the outcome status of the deploy run.