dsh-spec-gate
Run the following command in DeepSeek Harness:
dsh plugin install boomzikazita/dsh-spec-gate
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install boomzikazita/dsh-spec-gate in your terminal to install the dsh-spec-gate plugin. Source: https://github.com/boomzikazita/dsh-spec-gate
About this plugin
When you ask an AI to write a report or run an analysis, the brief is often one sentence long, yet the result still misses the mark. dsh-spec-gate takes a different approach: lock the specification before any work begins, then gate delivery against it, refusing to hand over anything that does not meet the locked criteria.
Two gates do the heavy lifting. The interview gate uses Socratic questioning to surface hidden assumptions, scoring eight dimensions (audience, purpose, format, data, scope, tone, acceptance criteria, deadline) and only proceeding once the ambiguity score converges at 0.2 or below; the result is sealed into an immutable, hash-keyed Seed via spec_seed_build. The acceptance gate then mechanically checks the draft against that Seed before delivery, testing keyword coverage, hard constraints, and word count, and fails closed on any miss.
It is built for content production workflows such as writing, analysis, and reports. It pairs naturally with dsh-cybernetics: every acceptance failure is logged to the events stream and fed back into the cybernetic control loop, giving the next iteration a concrete, auditable signal to work from.
Use Cases
- Clarify vague requests via Socratic questioning before any work begins
- Gate delivery with mechanical checks against a locked specification seed
- Log acceptance failures to drive cybernetic feedback iterations
Best For
- Producers of AI-written reports, analyses, and long-form content
- Teams with explicit acceptance criteria for deliverables
- Developers pairing with dsh-cybernetics for closed-loop iteration
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.