dsh-skill-authoring
Run the following command in DeepSeek Harness:
dsh plugin install boomzikazita/dsh-skill-authoring
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install boomzikazita/dsh-skill-authoring inside DeepSeek Harness to install the plugin (source: https://github.com/boomzikazita/dsh-skill-authoring).
About this plugin
As a DSH skill library grows, many agent skills end up shipped without a consistent skeleton: no clear trigger conditions, no boundary declarations, no quick-reference patterns, no end-to-end examples, and no verification trail. The result is unpredictable agent behavior and no shared quality bar across a team. dsh-skill-authoring exists to close that gap.
It is built around a nine-point quality gate covering naming convention, description precision, trigger conditions, boundaries, reusable quick references, worked examples, reference splitting, uncertainty annotation, and operational-manual writing style. Three tools sit behind the gate: skill_scaffold generates a compliant SKILL.md skeleton plus references/index.md in one step; skill_audit scores an existing file against all nine gates and prints a pass or fail table; skill_check_name validates the name against the DSH regex. A write-guard hook is also installed: any write or edit targeting a SKILL.md under your skills directory triggers an automatic audit and warns the agent of gates that have not yet been met, embedding quality control directly into the everyday editing workflow.
This plugin is aimed at developers and tech leads maintaining a DSH skill library, as well as agent engineers who need a repeatable standard for every new or existing skill, whether you are scaffolding your first one or running a compliance pass across dozens.
Use Cases
- Start new skills from a compliant skeleton to avoid structural gaps and naming violations
- Run a batch nine-gate audit on existing skills to spot weaknesses and close them systematically
- Enforce consistent naming, writing style, and example standards across the team to cut collaboration overhead
Best For
- Developers and platform engineers maintaining a DSH skill library
- Agent workflow leads who need a team-wide quality standard
- Tech leads who treat skill documentation as a gated, auditable artifact
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.