AI Agent Hub
Back to plugins
🖥️

dsh-strength-refine

Client Updated 2026.08.26

Run the following command in DeepSeek Harness:

dsh plugin install ShadowBruceMeaningLau/dsh-strength-refine

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

Install by running dsh plugin install ShadowBruceMeaningLau/dsh-strength-refine in your DeepSeek Harness terminal; the full source is at https://github.com/ShadowBruceMeaningLau/dsh-strength-refine .

About this plugin

A vague, high-level goal (mastering a domain, solving a cross-disciplinary problem, onboarding into a new field) usually ends in the soft verdict of "I think I know enough." dsh-strength-refine replaces that intuition with a verifiable artifact: through guided clarification and a layered gate workflow (clarify, then decompose, self-check, and approve level by level, then a final audit, and finally archiving), it produces a hierarchical capability-point tree where every leaf carries an acceptance criterion built from four elements (action, object, pass/fail criterion, and evidence) and a four-part proof (definition, necessity, coverage, boundary) demonstrating that the tree fully covers the original requirement with no true redundancy.

The internal logic of the tree is made explicit: leaves are linked by a prerequisite DAG, collaborative argument clusters, or alternative paths; cross-level proof chains are consolidated in a final audit; the archive includes a learning-path suggestion traced along the logic graph; and four Mermaid diagrams render natively in Obsidian. What you receive is not a to-do list but a capability proof, every node checkable, every edge accounted for.

Ideal for self-directed learners, knowledge workers, and technical leads who need to model, communicate, and validate competencies, especially anywhere the question shifts from "do I roughly get it?" to "can I pass the check on each capability point, right now?"

Use Cases

  • Break a vague goal like mastering data science into a verifiable capability checklist
  • Build a role competency model with actionable acceptance criteria for team onboarding
  • Generate a complete, redundancy-free capability coverage tree before starting a new domain

Best For

  • Self-directed learners who want to turn I roughly get it into I pass every capability check
  • Knowledge workers and team leads who need to model, communicate, and validate competencies
  • Project planners converting vague goals into discrete, verifiable capability units