dsh-plugin-autoqueue
Run the following command in DeepSeek Harness:
dsh plugin install alin-ever/dsh-plugin-autoqueue
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install alin-ever/dsh-plugin-autoqueue to install the plugin; its source is hosted at https://github.com/alin-ever/dsh-plugin-autoqueue.
About this plugin
Anyone who has pushed a batch of Markdown prompts through DSH probably knows the drill: paste the prompt into the foreground chat, watch it run, manually resume when it hits a block, retry on timeout, then clean up the output yourself. Once the queue grows, foreground conversation and background dispatch start fighting for the same session. autoqueue solves this with a deliberately narrow scope — an unattended task queue that dispatches, unblocks, retries, settles, and archives tasks on its own, while your foreground DSH sessions keep chatting exactly as if nothing changed.
The isolation model is session-granular. Every task binds to a dedicated session and its own working directory, with approvalPolicy pinned to never and execution restricted to two versioned engine presets that forbid questions, detached processes, and any escape from the owned session. When a foreground session is active, a running goal goes through a persistent pause-before-cancel hand-off and only resumes after two consecutive trusted idle observations — no duplicate prompt injection. On the outside, the plugin exposes 16 autoqueue_* Host tools, a full HTTP API with an OpenAPI 3.1 description, and a compact SSE event stream. The built-in React dashboard in the DSH sidebar covers the entire lifecycle from creation and editing to stopping, archiving, and restoring. Scheduling supports one-shot ISO 8601 timestamps, five-field cron recurrence, wall-clock deadlines, and per-attempt timeouts across a six-state machine.
autoqueue is the missing queue layer if you have structured Markdown tasks that need to run on a schedule, in a loop, or on demand from an external AI, CI pipeline, or script — without ever touching the foreground. It does not replace DSH; it simply gives DSH a pair of hands for the background.
Use Cases
- Drop a batch of Markdown tasks into the queue with cron or one-shot timestamps for recurring execution while foreground coding stays uninterrupted
- Dispatch tasks to DSH from CI scripts or external AIs over the HTTP API and pull execution reports without watching foreground sessions
- When a task hits a block, the engine auto-injects steering and resumes; on repeated unreachability it persists a cancel intent before retrying — no manual intervention needed
Best For
- Developers who batch-process structured Markdown tasks through DSH daily
- Anyone who needs DSH to run scheduled background jobs while the foreground stays responsive
- DevOps and automation engineers wiring DSH into external orchestration pipelines
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.