dsh-issue2pr
Run the following command in DeepSeek Harness:
dsh plugin install LONGSASASASASA/dsh-issue2pr
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install LONGSASASASASA/dsh-issue2pr in your DeepSeek Harness terminal to install the plugin; browse the source at https://github.com/LONGSASASASASA/dsh-issue2pr
About this plugin
Most Issue-to-PR coding agents chase a one-shot result, skipping understanding, planning, and verification. dsh-issue2pr takes the opposite approach: it decomposes the entire path into 11 non-skippable stages, each with numbered artifacts on disk, input/output contracts, and explicit failure signals. The deliverable is not just a diff but a full, auditable evidence chain from Issue to merged PR.
Core capabilities include multi-agent collaborative coding or delegation to Claude Code CLI and native DSH agents; a Patch Pipeline with a ledger that rolls back only agent-introduced changes; real test execution in a cloned repository with full output persisted for traceability; a three-dimensional review gate covering diff scope, API and security, and test reinforcement; and four human review checkpoints where you approve or reject with feedback. Every LLM call, git operation, and test run is logged in trace/events.jsonl, and a floating smart assistant lets you ask real-time questions about progress or failures.
It is built for teams and individual developers who want to bring AI-assisted coding under engineering control rather than treating it as a black box — especially those using the DSH host workflow, needing auditable PR descriptions and rollback-able patches, or preferring a staged delivery model that eases from manual gating toward full automation.
Screenshots
Use Cases
- Drive a GitHub Issue through to a reviewed, merged PR with every step producing a numbered artifact
- Require rollback-able patches and real test evidence instead of trusting a single opaque generation
- Phase your workflow from manual per-stage gating toward semi-automated pipeline delivery
Best For
- Teams using the DSH host that need engineering-grade control over AI-generated code
- Individual developers who want to move from black-box generation to auditable, traceable, rollback-able delivery
- Tech leads who need staged gatekeeping to reduce risk and gradually increase autonomy in AI-assisted coding
Related Plugins
A method pack that makes AI coding agents plan against your real baseline, prove completion with fresh evidence, and reduce reworks and unsafe changes.
Turns the DeepSeek Harness session into a captain that builds a durable sub-agent team, splits goals into dependency-aware tasks, and coordinates work via direct messages and a live Web UI.
Gives coding agents design judgment, letting Claude Code, Cursor, and 70+ agents generate and iterate high-quality UI, presentations, and graphics right from the terminal.
Run the Pi ecosystem's plugins on DeepSeek Harness, unmodified, via a compatibility layer that implements Pi's public extension ABI on DSH's native services.