Who Is Actor

- Contributor display names are now never included in output; names are used only locally to calculate aggregates and are discarded before leaving the host. - Updated documentation to clarify that no per-contributor results or author names are ever produced. - Further restricted any personnel-style analysis or individualized judgments, emphasizing exclusive support for aggregate repository metrics. - Removed legacy file skill-card.md to streamline documentation. - Minor updates to activation and privacy instructions to reinforce scope and usage boundaries.

- Tightened privacy policy: now clarifies that no per-contributor breakdown, ranking, or scoring is ever produced. - Updated usage notice to emphasize refusal of all personnel-style assessments or individualized reports, even if a repository path is provided. - Strengthened warnings: output is strictly repository-wide aggregates, not for HR or performance uses. - Documentation and activation phrase examples adjusted for greater clarity and explicitness regarding scope and ethical constraints. - Removed obsolete documentation file `skill-card.md`.

who-is-actor 1.0.13 - Refined all documentation to clarify that this skill provides only repository-level, aggregate collaboration-pattern analytics, not individualized developer profiling. - Updated activation phrases in SKILL.md to explicitly disallow and refuse requests for individual ranking, performance assessment, or personnel-style output. - Removed the file skill-card.md. - Revised the description and usage guidance in all README and intro files to emphasize the aggregate, non-personal nature of analysis. - Expanded and clarified privacy and consent instructions, including explicit guidance for refusal scenarios.

**Major update: Tightened privacy, consent, and data-handling rules; clarified skill scope and usage activation requirements.** - Agents must require explicit user consent, a concrete repository path, and authority confirmation before any analysis. - Skill can no longer be used for HR, performance, or personnel decisions; developer metrics are strictly technical and aggregate. - Only locally aggregated, redacted statistics are sent to the AI model—raw commit message text and file paths must never be included in prompts. - Enhanced activation protocol: general mentions of repository analysis don't trigger the skill—user must clearly state intent and repo path. - Documentation now outlines strict privacy, sensitive-data filtering, and user notification obligations before running commands.

who-is-actor v1.0.10 - Updated and refined documentation in SKILL.md, README.md, and SKILL_CN.md - No code logic or execution changes; this release is documentation-focused - Examples and multilingual prompts clarified for skill invocation - Security, parameter validation, and privacy notes remain unchanged

who-is-actor v1.0.9 - Documentation updates, including SKILL.md, README.md, and SKILL_CN.md. - Clarified skill description, usage instructions, input validation, and command safety in documentation. - No changes to the core logic or feature set; this is a documentation-focused update.

# who-is-actor v1.0.8 Changelog - Clarified "zero dependencies" to specify that standard Unix tools (`awk`, `sort`, etc.) are required but nothing is installed. - Added explicit note and section for a dry-run mode: users can preview all shell commands before execution for security review. - Documentation now directly links to the project repository. - Improved and reorganized security and privacy sections; all command validation and whitelist rules now highlighted and easier to follow. - Minor edits for clarity, conciseness, and improved user onboarding instructions.

who-is-actor 1.0.7 - Added SKILL_CN.md for Chinese-language documentation. - Expanded SKILL.md with natural language usage examples in multiple languages (English, 中文, 日本語, 한국어, Español, Français, Deutsch). - Clarified that natural language prompts are recommended and listed common sample queries. - No changes to core logic, security model, or feature set.

- Documentation and metadata kept up to date; no functional changes in this version. - Ensures all descriptions and instructions reflect the current design and behavior.

No user-facing changes detected in this version. - No file changes present; documentation and functionality remain unchanged.

who-is-actor 1.0.4 - Substantially improved English documentation in SKILL.md, providing a concise quick start, clearer parameter summary, and usage guidance for international users. - Added a highly readable Markdown overview to SKILL.md, making the skill’s purpose, parameters, workflow, security, and privacy rules more accessible. - Tightened various descriptions: clarified validation rules, reinforced security specifications, and improved step-by-step instructions. - README.md unchanged in this release.

who-is-actor v1.0.3 - Tightened privacy and security controls: authors can only be specified via display names (no `@` or email allowed). - Updated input validation: author names must not contain `@` and now match regex `^[a-zA-Z0-9 _.-]+$`. - Clarified that the skill never accepts email addresses or exposes them in reports or via git commands. - Updated documentation for stricter rules and improved privacy instructions.

- Added an explicit shell command whitelist: only specific read-only git commands and basic text filters are allowed, all write/network/non-git shell operations are strictly forbidden. - Updated the security section to detail the allowed/forbidden commands, making it clearer and stricter on shell safety. - Added a usage disclaimer to the Engagement Index: it must not be used for HR decisions such as performance reviews or layoffs. - Improved documentation for security best practices and ethical responsibility when analyzing developer activity. - No changes to core logic/command usage—changes are documentation and security policy clarifications only.

**Summary:** This version adds strict command injection prevention and privacy rules. The "Slacking Index" is replaced with a neutral "Engagement Index". - All shell command parameters (repo path, author, dates, branch) must pass new strict validation rules before use, preventing command injection. - Privacy protections: developer emails are no longer collected or shown; only author names are processed and reported. - "Slacking Index" is replaced by "Engagement Index"—now a neutral measure of visible contribution, with clear warnings about its meaning and limitations. - New documentation sections on input validation rules, privacy rules, and security-first principles. - Updated output structure and field names, reflecting the terminology changes (e.g., "slacking" replaced by "engagement"). - Expanded contextual warnings in report interpretations regarding what Git data can and cannot reflect.

- Initial release of the "who-is-actor" skill: zero-dependency, no-script Git repository developer profiling. - Provides comprehensive reports on each developer's commit habits, work habits, efficiency, code style, code quality, and slacking index—entirely from native git CLI output analyzed by AI. - Output includes detailed per-developer dashboards, AI evaluations, improvement suggestions, radar scores, and one-line summaries. - Supports team-wide comparison, health evaluation, and bus factor risk detection. - Strictly uses only native git commands for data collection; no external dependencies required.