| name | algorithm-tutorial-deepen |
| description | v0.1.0 - Deepen 1-2 selected concepts in a reviewed algorithm tutorial using PDKH standards. Use when a method tutorial needs stronger invariant, formalization, correctness, complexity, counterexample, or engineering-reality evidence without broadening into unrelated topics. |
Algorithm Tutorial Deepen
Overview
Deepen selected concepts after the tutorial structure is in place. This skill
does not add broad new topics. It applies the PDKH ladder to 1-2 concepts so
the article becomes more rigorous.
When to Use
- A plan selected concepts for deepening.
- Review passed structure but requested more depth.
- The tutorial needs stronger invariants, formulas, correctness, complexity,
counterexamples, or engineering caveats.
When NOT to use: first drafting, review acceptance, broad expansion,
LeetCode problem optimization, or SEO enhancement.
Reference Map
../references/deepening-ladder.md
../references/depth-checklist.md
The Deepening Loop
- Select Concepts
- Use only 1-2 concepts named by the plan or review.
- Apply PDKH
- Problem reframe.
- Minimal worked example.
- Invariant or contract.
- Formalization.
- Correctness sketch.
- Thresholds and complexity.
- Counterexample or failure mode.
- Engineering reality.
- Keep Scope Narrow
- Add evidence to existing sections or one targeted deepening section.
- Do not create unrelated top-level topics.
- Report Evidence
- Name which PDKH steps were satisfied.
Output Format
## Deepening Result
- status: needs_review
- concepts:
- PDKH evidence:
## Added Or Revised Content
...
## Review Handoff
- review_skill: algorithm-tutorial-review
Guardrails
- Do not broaden the article to meet length.
- Do not add generic background.
- Do not invent claims or benchmarks.
- Do not skip counterexamples when the concept has common failure modes.
Verification