Code Analysis Skills

Republish of v1.1.0 with a corrected pure-semver version string. ClawHub UI was rendering vv1.1.0 because the previous publish included a leading v. No code changes vs v1.1.0.

Defence-in-depth + doc cleanup: replaced misleading -a Your-Name README and SKILL examples (the flag was removed in v1.0.7) with the real --multi-author-team-retro + --consented-author flow; dropped weekday-distribution fields friday_ratio and monday_ratio from cadence_signal_analyzer to remove the late-week-skew repackaging path; reframed every analyzer docstring from per-author to per-consented-identity and clarified that BaseAnalyzer authors arg is an internal consent-gated channel populated only by src.main; documented that scanner.py performs filesystem enumeration only and that --scan-all does not widen people-data exposure because each repo still passes through the consent gate; ReflectionNarrator.narrate now accepts a consented_identities allow-list and silently drops any unconsented identity that slipped past the upstream gate.

Doc/code alignment + dead-code purge: removed -a/--author from all docs (CLI never had it after v1.0.7); documented --multi-author-team-retro and --consented-author in every README/SKILL table; updated project tree to point at cadence_signal_analyzer.py and narrator/reflection_narrator.py; deleted unused *_score intermediates (sparsity_score / trivial_score / non_code_score / late_week_skew_score / add_delete_imbalance_score / disappearance_score / low_output_score) from cadence analyzer; removed legacy 'formerly DeveloperEvaluator' callout from narrator header.

Structural safety overhaul: self-scope by default; cross-author analysis requires --multi-author-team-retro plus per-person --consented-author; removed env-var bypass for the consent gate; dropped composite 0-100 score, S/A/B/C/D/E/F letter band, verdict line, leaderboard and cross-author comparison tables from every output format; dropped pylint dependency; removed legacy report.md sample.

Reframed as a Git-history reflection tool: descriptive metrics only, removed people-ranking framing, added mandatory usage-policy gate (--i-have-consent), neutralized report wording, replaced leaderboards with alphabetical overviews.

Version 1.0.6 - Added a direct link to the project's GitHub repository at the top of documentation files. - Improved clarity and navigation in documentation for easier user access. - No functional code or API changes; updates are limited to documentation improvements.

- Added multilingual natural language examples (English, 中文, 日本語, 한국어, Español, Français, Deutsch) for easier usage and discovery. - Improved usage guidance: clarified natural language triggering as the recommended method. - Updated documentation structure for better readability and quick reference. - No changes to analysis logic or CLI features; only documentation and usage instructions were updated.

**Skill version 1.0.4 changelog** - Added a new section describing natural language skill plugin usage, enabling analysis via conversation without CLI commands. - Updated documentation to highlight "Use as a Skill Plugin" for user ease. - Clarified that the skill can now identify repository and run analysis automatically in conversational contexts. - No changes to functionality or parameters. Documentation enhanced for better usability.

**Changelog for code-analysis-skills v1.0.3** - Added Chinese-language skill documentation (SKILL.cn.md) for improved accessibility. - Rewrote SKILL.md in concise English with clearer structure, usage, and parameters. - Added Quick Start, CLI parameter reference table, and streamlined workflow instructions to SKILL.md. - Improved emphasis on privacy and appropriate usage in documentation. - No changes to core functionality—documentation only.

**Major update introducing developer evaluation, "slacking index", and PDF reporting.** - Added slacking analyzer to assess developer "摸鱼指数" (slacking index) and related behavioral patterns. - Introduced comprehensive developer evaluation with direct scoring, strengths, weaknesses, and actionable suggestions. - Expanded output formats to support PDF reports, in addition to Markdown, HTML, and JSON. - Enhanced report content: now includes developer scorecards, rankings, and detailed explanations for each evaluation. - Updated documentation to reflect new usage scenarios, privacy guidance, evaluation criteria, and detailed output samples. - Added new modules for slacking analysis, scoring logic, and PDF generation in the codebase.

- Added a new Chinese-language README file (README.CN.md) for improved accessibility. - Updated README.md with minor revisions (no functional changes to the skill itself).

Initial release of code-analysis skill. - Analyze and compare Git repositories for commit habits, work routines, development efficiency, code style, and code quality. - Supports scanning a single repo or all repos in a directory, filtering by author, date range, and branch. - Generates structured reports in markdown (default), JSON, or HTML formats. - Provides detailed findings and comparisons across five dimensions: commit habits, work habits, efficiency, code style, and quality. - Includes command-line scripts and modular analyzers for flexible usage. - Reference guide available for metric definitions and interpretations.