davidamitchell/Skills

adversarial-review

Runs a structured adversarial critique process using opposing reviewer roles and synthesis to expose hidden risks, weak reasoning, and blind spots in code, plans, documents, strategies, or agent outputs.

原文语言：英语

code-review

Performs systematic, multi-dimensional code review covering correctness, security, performance, maintainability, and style. Produces actionable, prioritised findings with evidence-based recommendations. Use when asked to review code, audit a pull request, or…

原文语言：英语

backlog-manager

Manages an outcome-focused backlog stored in /work/backlog.md. Handles command-driven operations to add, refine, list, start, complete, and archive work items. Each item is tracked with a status, creation date, and a clear observable outcome statement. Use…

原文语言：英语

backlog-worker

Executes backlog items one at a time — selects the next ready item, decomposes it into atomic actions, performs each action using the appropriate sub-skill, reviews the output, records learnings, and advances the item to done. Use when asked to work the…

原文语言：英语

swe

Applies SOLID principles, Fielding's REST architectural constraints, Gang of Four design patterns, and Enterprise Integration Patterns to produce well-designed, maintainable software. Use when designing systems, implementing features, making architectural…

原文语言：英语

skill-author

Authors a new SKILL.md file from scratch; scans the existing skill library for related skills before drafting to prevent duplication and ensure interoperability, then audits those related skills after drafting to add any required cross-references. Use when…

原文语言：英语

citation-discipline

Ensures all factual statements are bound to verifiable sources, cited precisely at the point of claim, and integrated without disrupting prose flow. Use when producing research, reports, or any writing where every assertion must be independently traceable.

原文语言：英语

inline-citation

Formats hyperlinked inline citations as Author (Year) anchors placed immediately after the claim they support. Use when writing blog posts, technical documentation, internal knowledge bases, or web-published research summaries that require visible attribution…

原文语言：英语

tdd

Guides implementation of any feature or fix using test-driven development; use when writing new production code, fixing bugs, or changing existing behaviour.

原文语言：英语

remove-ai-slop

Removes generic, formulaic patterns from AI-generated text to produce more natural, authentic prose. Targets predictable phrases, structural uniformity, false agency, alignment artifacts, and lexical monotony. Use when AI-drafted content feels flat,…

原文语言：英语

adr

Creates and maintains Architecture Decision Records (ADRs) — structured documents that capture the context, decision, consequences, and alternatives for significant architectural choices. Use when recording a new technical decision, updating the status of an…

原文语言：英语

research

Performs rigorous, evidence-driven research using recursive decomposition, iterative verification loops, and disciplined reasoning. Use when asked to research a topic, investigate a question, or gather and synthesize evidence into structured findings.

原文语言：英语

speculation-control

Enforces strict separation between evidence-based statements and speculative, interpretive, or subjective content. All non-factual content must be explicitly labeled and grounded in known information. Use when producing research, analysis, or factual writing…

原文语言：英语

peer-reviewer

Checks a completed research item for logical coherence, alternative explanations, and cross-item integration. Use after research investigation is complete and before the item is marked done. Audit-only — never edits the item.

原文语言：英语

research-reviewer

Audits a completed research item for citation discipline, speculation control, writing quality, and logical coherence. Use after research investigation is complete and before the item is marked done. Audit-only — never edits the item.

原文语言：英语

research-question

Validates and scopes a research question before investigation begins. Use when formulating a new research item, reviewing a drafted question for quality, or checking whether a question is ready to hand to the research skill.

原文语言：英语

strategy-author

Transforms context into high-quality strategy using Rumelt's Kernel (Diagnosis, Guiding Policy, Coherent Actions) and Porter's trade-off principles. Use when producing strategy documents, evaluating strategic options, or reviewing and critiquing existing…

原文语言：英语

feedback

Delivers structured, constructive critique of written work, arguments, decisions, or plans. Findings are specific, evidence-grounded, and paired with actionable recommendations. Use when asked to review, critique, assess, or improve any piece of work where…

原文语言：英语

plain-language

Rewrites complex, technical, or dense text so that a non-expert reader can understand it without losing accuracy or completeness. Use when asked to simplify content, make text more accessible, or adapt specialist material for a general audience.

原文语言：英语

technical-writer

Produces clear, accurate technical documentation structured for its intended audience — developers, operators, or end users. Covers READMEs, API references, guides, runbooks, and architecture documents. Use when asked to write, audit, or improve technical…

原文语言：英语

strategic-persuasion

Constructs clear, well-structured persuasive arguments grounded in audience values, sound reasoning, and precise language. Use when building persuasive content, strategic communications, or audience-targeted messaging that must achieve a specific belief or…

原文语言：英语