AI Agent Hub
Back to plugins
⚙️

dsh-product-delivery-workflow

Workflow Updated 2026.08.15

Run the following command in DeepSeek Harness:

dsh plugin install wellorbetter/dsh-product-delivery-workflow

Paste the following prompt into your AI chat to install this plugin:

Run dsh plugin install wellorbetter/dsh-product-delivery-workflow in DeepSeek Harness to install this plugin; the source repository is https://github.com/wellorbetter/dsh-product-delivery-workflow , and a profile restart is required to activate it.

About this plugin

Taking a product from a brief to a full release usually means stitching together research, architecture, implementation, multi-round code review, automated testing, and a pre-release audit. Doing all of that by hand is slow and tends to stall at the review-and-fix loop. dsh-product-delivery-workflow packs the entire pipeline into one autonomous overnight run: hand it a brief before bed, and the orchestrator walks through ten stage gates, routes every review finding back to the responsible agent, re-reviews automatically for up to five cycles, and produces a final report with a full artifact trail by the time you wake up.

The core capability is a ten-stage, artifact-driven pipeline with true loop closure. One orchestrator skill coordinates nine role agents-product manager, senior architect, senior developer, multi-agent swarm, performance reviewer, test agent, memory curator, evolution agent, and release audit agent-each working inside explicit contracts with disjoint file ownership. Parallel waves keep throughput high while safety rails guarantee the release audit fires before any commit, PR, or push, and memory curation never stores secrets or raw transcripts. Every artifact lands in a local directory so the entire run is auditable after the fact, with nothing claimed unless its verification result actually exists.

It is built for solo developers and small teams who want to delegate a complete delivery cycle to AI without babysitting each session. There is no build step and no extra configuration; after installing the plugin into a target profile, simply ask for the workflow in any session and let the ten-stage loop run to completion.

Use Cases

  • Kick off a full product-to-release pipeline before bed and review the final report and artifact trail in the morning
  • Run autonomous implementation, parallel review, and up-to-five fix cycles against an existing PRD until all gates pass
  • Trigger a pre-release security audit covering secrets, privacy, and code hygiene before any commit or PR is pushed

Best For

  • Solo developers who want to hand off an entire delivery cycle to AI without babysitting each session
  • Small teams that need multi-round review closure without manually tracking every fix iteration
  • AI workflow practitioners who value end-to-end unattended runs with full artifact auditability