Skip to main content

lexical-analysis

Prompt for Lexical Larry to perform a lexical style diagnostic and minimal diff pass

来源信息

仓库
sddevelopment-be/quickstart_agent-augmented-development
最近来源活动
2026年1月31日 06:30
检测到的 SKILL.md 语言
英语
星标
13
Forks
3

安装方式

默认使用会先检查来源的 Prompt;你也可以切换为直接命令,或下载本地副本。

检查来源文件

决定是否安装前,请先阅读 SKILL.md,以及 SkillsMP 当前展示的配套文件。

正在显示 SKILL.md

SKILL.md
来源说明 · 只读预览
name
LEXICAL ANALYSIS
description
Prompt for Lexical Larry to perform a lexical style diagnostic and minimal diff pass
version
1.0.0
type
prompt-template
agent
lexical-larry
category
documentation
tags
["lexical","style","tone","diagnostics","minimal-diffs"]
# LEXICAL ANALYSIS Prompt for Lexical Larry to perform a lexical style diagnostic and minimal diff pass ## Instructions Clear context. Bootstrap as Lexical Larry. When ready: Perform lexical analysis + minimal diff proposal. ## Inputs: - Target Files (comma paths): \<FILES> - Style Anchor (path or none): \<ANCHOR> - Directives (comma codes, optional): \<DIRECTIVES> - Tone Objectives (bullets): \<TONE> - Diff Strictness (minimal|moderate|advisory): \<STRICTNESS> - Report Depth (summary|full): \<DEPTH> - Include Before/After Snippets? (yes/no): \<SNIPPETS> ## Task: 1. Load style baseline from \<ANCHOR> or `docs/specific_guidelines.md`. 2. Generate `work/LEX/LEX_REPORT.md` (sections: Tone, Rhythm, Markdown Hygiene, Anti-Fluff, Medium Fit, Clarity Before Complexity). 3. Produce `work/LEX/LEX_DELTAS.md` with diff groups per violated rule; obey \<STRICTNESS>. 4. If \<SNIPPETS> = yes: include before/after blocks for non-trivial changes. 5. Provide confidence markers (✓ / ⚠ / ❗) per rule. 6. Summarize metrics: avg sentence length, para length variance, modality issues. ## Output: - LEX_REPORT.md content - LEX_DELTAS.md content - Metrics summary block ## Constraints: - Preserve author voice; avoid reflow unless necessary. - No hype, flattery, or style homogenization. - Only modify lines with concrete rule justification. Ask clarifying questions if \<FILES> missing. ## Workflow 1. Summarize metrics: avg sentence length, para length variance, modality issues. ## Inputs - **FILES** (required) - comma paths - **ANCHOR** (required) - path or none - **DIRECTIVES** (required) - comma codes, optional - **TONE** (required) - bullets - **STRICTNESS** (required) - minimal|moderate|advisory - **DEPTH** (required) - summary|full - **SNIPPETS** (required) - yes/no ## Outputs - Metrics summary block
在 GitHub 查看