feat!: rebrand to @devran-ai/kit v4.0.0 and optimize framework tokens#2
feat!: rebrand to @devran-ai/kit v4.0.0 and optimize framework tokens#2emredursun merged 4 commits intodevfrom
Conversation
- Untrack 917 node_modules files (were committed before .gitignore rule) - Delete npm-publish-output.txt build artifact - Add npm-publish-output.txt to .gitignore - CI should run npm install instead of relying on committed deps
BREAKING CHANGES: - Package renamed: antigravity-ai-kit → @devran-ai/kit - CLI command renamed: ag-kit → kit - Scaffolder renamed: create-antigravity-app → create-kit-app Rebranding (127 files): - Full rename across package.json, CLI, lib/, docs/, .agent/, tests/ - README rewritten for enterprise clarity (758 → 120 lines) - Removed node_modules from git tracking (917 files, 236K lines) - Deleted documentation bloat (archive docs, operational guides) Framework token optimization (56% reduction, 262K chars saved): - Workflows: 125K → 63K (50% reduction, all under 10K chars) - Agents: 273K → 114K (58% reduction, all under 10K chars) - Skills: 78K → 26K (67% reduction for top 5 offenders) - Extracted shared rules to workflow-standards.md - Removed textbook content AI already knows - Deduplicated governance sections across 22 workflows All 348 tests passing.
Summary of ChangesHello, I'm Gemini Code Assist1! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed! This pull request delivers a comprehensive overhaul of the AI development framework, focusing on a complete brand identity shift and substantial performance enhancements. The changes aim to modernize the project's naming conventions and significantly reduce its footprint and token consumption, thereby improving efficiency and maintainability. Highlights
Ignored Files
Using Gemini Code AssistThe full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips. Invoking Gemini You can request assistance from Gemini at any point by creating a comment using either
Customization To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a Limitations & Feedback Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for GitHub and other Google products, sign up here. Footnotes
|
|
C:/Program Files/Git/gemini review |
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
There was a problem hiding this comment.
Code Review
The pull request primarily focuses on a comprehensive rebranding of the 'Antigravity AI Kit' to 'Devran AI Kit', updating all relevant files including documentation, agent definitions, CLI commands, and internal module references. This also involves updating the kit version to 4.0.0 and introducing a new workflow-standards.md rule. Review comments highlight critical markdown rendering issues in the report formats for the explorer agent, planner, and PR reviewer, where closing backticks are missing. Additionally, there are suggestions to standardize agent titles by explicitly including 'Devran AI Kit' for consistency across all agent files. A changelog entry's clarity regarding the removal versus ignoring of npm-publish-output.txt is also noted as an area for improvement.
| ## Technical Debt Inventory (table: category, count, severity, priority) | ||
| ## Metrics (table: metric, value, assessment) | ||
| ## Recommendations (prioritized: CRITICAL → HIGH → MEDIUM) | ||
| ``` |
There was a problem hiding this comment.
The markdown code block for the report format is not closed. It's missing the closing backticks, which will break rendering of this section.
| ``` | |
| ## Recommendations (prioritized: CRITICAL → HIGH → MEDIUM) |
There was a problem hiding this comment.
Verified: code block is properly closed at line 123. The opening backticks are at line 115 and closing at line 123. This is a false positive.
|
|
||
| Proceed with this plan? (yes / no / modify) | ||
| **WAITING FOR CONFIRMATION** — Proceed? (yes / no / modify) | ||
| ``` |
There was a problem hiding this comment.
The markdown code block for the implementation plan template is not closed. It's missing the closing backticks, which will break rendering of this section.
| ``` | |
| **WAITING FOR CONFIRMATION** — Proceed? (yes / no / modify) |
There was a problem hiding this comment.
Verified: code block opens at line 89 and closes at line 111. Properly matched. False positive.
| ## Verdict: {REQUEST_CHANGES | APPROVE | COMMENT} | ||
|
|
||
| {1-2 sentence justification referencing specific findings} | ||
| ``` |
There was a problem hiding this comment.
The markdown code block for the review output format is not closed. It's missing the closing backticks, which will break rendering of this section.
| ``` | |
| ## Verdict: {REQUEST_CHANGES | APPROVE | COMMENT} |
There was a problem hiding this comment.
Verified: code block opens at line 89 and closes at line 115. Properly matched. False positive.
| @@ -11,336 +11,128 @@ relatedWorkflows: [orchestrate] | |||
|
|
|||
| # Database Architect | |||
There was a problem hiding this comment.
For consistency with other agent files that were rebranded, consider adding the kit name to the title. This helps maintain a uniform structure across all agent definitions.
| # Database Architect | |
| # Devran AI Kit — Database Architect |
There was a problem hiding this comment.
Addressed in c420bcf — standardized all agent titles to clean format without prefix (e.g. '# Database Architect' instead of '# Devran AI Kit — Database Architect'). The prefix adds no value as the AI knows which project it's in.
| @@ -11,587 +11,134 @@ relatedWorkflows: [orchestrate] | |||
|
|
|||
| # Senior Staff DevOps Engineer | |||
There was a problem hiding this comment.
For consistency with other agent files that were rebranded, consider adding the kit name to the title. This helps maintain a uniform structure across all agent definitions.
| # Senior Staff DevOps Engineer | |
| # Devran AI Kit — Senior Staff DevOps Engineer |
There was a problem hiding this comment.
Addressed in c420bcf — standardized all agent titles to clean format without prefix (e.g. '# Database Architect' instead of '# Devran AI Kit — Database Architect'). The prefix adds no value as the AI knows which project it's in.
| @@ -10,348 +10,135 @@ relatedWorkflows: [orchestrate, ui-ux-pro-max] | |||
|
|
|||
| # Senior Frontend Architect | |||
There was a problem hiding this comment.
For consistency with other agent files that were rebranded, consider adding the kit name to the title. This helps maintain a uniform structure across all agent definitions.
| # Senior Frontend Architect | |
| # Devran AI Kit — Senior Frontend Architect |
There was a problem hiding this comment.
Addressed in c420bcf — standardized all agent titles to clean format without prefix (e.g. '# Database Architect' instead of '# Devran AI Kit — Database Architect'). The prefix adds no value as the AI knows which project it's in.
| @@ -11,289 +11,140 @@ relatedWorkflows: [orchestrate] | |||
|
|
|||
| # Mobile Developer | |||
There was a problem hiding this comment.
For consistency with other agent files that were rebranded, consider adding the kit name to the title. This helps maintain a uniform structure across all agent definitions.
| # Mobile Developer | |
| # Devran AI Kit — Mobile Developer |
There was a problem hiding this comment.
Addressed in c420bcf — standardized all agent titles to clean format without prefix (e.g. '# Database Architect' instead of '# Devran AI Kit — Database Architect'). The prefix adds no value as the AI knows which project it's in.
| @@ -11,528 +11,165 @@ relatedWorkflows: [orchestrate] | |||
|
|
|||
| # Performance Optimizer | |||
There was a problem hiding this comment.
For consistency with other agent files that were rebranded, consider adding the kit name to the title. This helps maintain a uniform structure across all agent definitions.
| # Performance Optimizer | |
| # Devran AI Kit — Performance Optimizer |
There was a problem hiding this comment.
Addressed in c420bcf — standardized all agent titles to clean format without prefix (e.g. '# Database Architect' instead of '# Devran AI Kit — Database Architect'). The prefix adds no value as the AI knows which project it's in.
| @@ -9,381 +9,133 @@ relatedWorkflows: [orchestrate] | |||
|
|
|||
| # Security Reviewer Agent | |||
There was a problem hiding this comment.
For consistency with other agent files that were rebranded, consider adding the kit name to the title. This helps maintain a uniform structure across all agent definitions.
| # Security Reviewer Agent | |
| # Devran AI Kit — Security Reviewer Agent |
There was a problem hiding this comment.
Addressed in c420bcf — standardized all agent titles to clean format without prefix (e.g. '# Database Architect' instead of '# Devran AI Kit — Database Architect'). The prefix adds no value as the AI knows which project it's in.
CHANGELOG.md
Outdated
| - README rewritten for enterprise clarity (758 lines to ~120 lines) | ||
| - Removed `node_modules/` from git tracking (917 files, 236K lines) | ||
| - Removed documentation bloat (archive docs, operational guides) | ||
| - Removed `npm-publish-output.txt` build artifact |
There was a problem hiding this comment.
This changelog entry is a bit confusing. The .gitignore file is being updated to ignore npm-publish-output.txt, while this note says the artifact is removed. If the intent is to stop tracking this file, the .gitignore change is correct. Perhaps rephrasing this to 'Stopped tracking npm-publish-output.txt build artifact' would be clearer.
There was a problem hiding this comment.
Fixed in c420bcf — reworded to 'Untracked node_modules/ from git — added to .gitignore' and 'Deleted npm-publish-output.txt build artifact, added to .gitignore'.
|
It looks like you tried to invoke the |
- Fix 3 frontmatter descriptions exceeding 250-char limit - Standardize agent titles: remove redundant "Devran AI Kit —" prefix - Fix CHANGELOG wording: clarify node_modules untracking vs deletion - Add token optimization entry to CHANGELOG
|
C:/Program Files/Git/gemini review |
|
/gemini review |
There was a problem hiding this comment.
Code Review
This pull request is a significant and well-executed effort to rebrand the project from antigravity-ai-kit to @devran-ai/kit. The changes are comprehensive, covering package names, CLI commands, documentation, and internal file content. In addition to the rebranding, the PR introduces substantial token optimization across agent, skill, and workflow files by making them more concise and extracting shared rules into a new workflow-standards.md file. This is a smart refactoring that improves maintainability and efficiency. The cleanup of the git repository by untracking node_modules is also a welcome improvement. The complete rewrite of the main README.md file makes the project much more approachable for new users. Overall, the changes are consistent, well-documented in the PR description, and align perfectly with the stated goals.
The ASCII diagram header still showed v3.8.0 after the rebrand.
* feat(gitflow): add GitFlow branching strategy with dev integration branch - Create dev branch as integration target for all feature work - Update CI pipeline to run on both main and dev branches - Add PR template with conventional commit checklist - Rewrite git-workflow rules with full GitFlow branch strategy - Update CONTRIBUTING.md with GitFlow workflow documentation - Update /pr workflow to auto-detect GitFlow and target dev * chore: migrate repository URLs from besync-labs to devran-ai/kit Update all repository references across 17 files: - package.json, manifest.json: repository URL - README, CONTRIBUTING, docs: GitHub links - mkdocs.yml, robots.txt: GitHub Pages URLs - bin/ag-kit.js, create-antigravity-app: CLI output URLs * fix: address gemini-code-assist review findings on PR #1 - Fix HIGH: docs/getting-started.md cp command references old dir name - Fix MEDIUM: git-workflow.md hotfix merge instruction clarified - Fix MEDIUM: CONTRIBUTING.md restore skill/workflow frontmatter examples * chore: remove node_modules from git tracking - Untrack 917 node_modules files (were committed before .gitignore rule) - Delete npm-publish-output.txt build artifact - Add npm-publish-output.txt to .gitignore - CI should run npm install instead of relying on committed deps * feat!: rebrand to @devran-ai/kit v4.0.0 and optimize framework tokens BREAKING CHANGES: - Package renamed: antigravity-ai-kit → @devran-ai/kit - CLI command renamed: ag-kit → kit - Scaffolder renamed: create-antigravity-app → create-kit-app Rebranding (127 files): - Full rename across package.json, CLI, lib/, docs/, .agent/, tests/ - README rewritten for enterprise clarity (758 → 120 lines) - Removed node_modules from git tracking (917 files, 236K lines) - Deleted documentation bloat (archive docs, operational guides) Framework token optimization (56% reduction, 262K chars saved): - Workflows: 125K → 63K (50% reduction, all under 10K chars) - Agents: 273K → 114K (58% reduction, all under 10K chars) - Skills: 78K → 26K (67% reduction for top 5 offenders) - Extracted shared rules to workflow-standards.md - Removed textbook content AI already knows - Deduplicated governance sections across 22 workflows All 348 tests passing. * fix: address gemini-code-assist review findings on PR #2 - Fix 3 frontmatter descriptions exceeding 250-char limit - Standardize agent titles: remove redundant "Devran AI Kit —" prefix - Fix CHANGELOG wording: clarify node_modules untracking vs deletion - Add token optimization entry to CHANGELOG * fix: update architecture diagram version to v4.0.0 The ASCII diagram header still showed v3.8.0 after the rebrand. * feat: Devran AI Kit v4.1.0 — cross-IDE support, multi-language reviewers, enhanced DX * feat: Devran AI Kit v4.1.0 — cross-IDE support, multi-language reviewers, enhanced DX Phase 1 — Cross-IDE Config Generation: - New lib/ide-generator.js (zero-dep, pure functions, transactional writes) - kit init generates Cursor (.mdc), OpenCode (.json), Codex (.toml) configs - kit update regenerates IDE configs non-destructively - create-kit-app scaffolder includes IDE configs - ADR-002 documents architectural decision - 18 new tests (path traversal, secret detection, format validation) Phase 2 — Multi-Language Reviewers: - typescript-reviewer.md (strict mode, no any, generics, satisfies) - python-reviewer.md (PEP 8, mypy strict, async, Pydantic) - go-reviewer.md (error wrapping, goroutine leaks, context.Context) - Count synced across 8+ files (manifest, README, CheatSheet, docs, loading-rules) - Orchestrate dynamicAgentPool updated with new reviewers Phase 3 — Enterprise README: - Full rewrite with comparison table, badges, CLI reference - "Why Devran AI Kit?" differentiation section - Cross-IDE support table, v4.1.0 highlights Phase 4 — GitHub Community: - FUNDING.yml, CODE_OF_CONDUCT.md (Contributor Covenant v2.1) - Enhanced PR template (IDE impact, manifest sync checkboxes) - Enhanced issue templates (IDE dropdown, version field) Phase 5 — Enhanced Continuous Learning: - New lib/learning-engine.js (confidence scoring, clustering, decay) - Netflix-inspired feedback loop: observe → score → cluster → promote → decay - 16 new tests with input validation hardening Phase 6 — MCP Server Templates: - 5 pre-configured templates (GitHub, Supabase, Vercel, Filesystem, PostgreSQL) - ${VAR} placeholders for secrets (no hardcoded values) - IDE generator transforms templates into IDE-native MCP format Phase 7 — Documentation: - docs/faq.md, docs/cli-reference.md, docs/cross-ide-setup.md - Enhanced landing page with v4.1.0 highlights - mkdocs.yml nav updated Security hardening (from Tier-1 audit): - Null-byte rejection in path validation (H-1) - Input type/range validation on all learning-engine functions (H-2) - Parameter type validation on writeIdeConfigs (M-3) Verification: 382/382 tests passing, kit verify 127/0, kit scan 0 critical/0 high * fix: address Tier-1 code review findings (3 HIGH, 3 MEDIUM) HIGH-1: Use CURSOR_DIR/OPENCODE_DIR/CODEX_DIR constants instead of hardcoded path strings in ide-generator.js — prevents drift. HIGH-2: Extract atomicWriteFile() and rollbackWrittenFiles() helpers from writeIdeConfigs — each function now under 50 lines. HIGH-3: Fix TOML serializer for mixed scalar/object sections — emit [section] header once for scalars, then subsections separately. MEDIUM-4: Remove dead code — duplicate skip logic in writeIdeConfigs. MEDIUM-5: Freeze module.exports with Object.freeze for consistency. H-2 hardening: Add Array.isArray guards + Number.isFinite validation to learning-engine.js clusterPatterns and decayPatterns. * docs: add safety guarantees, agents listing, operating constraints to README Restores critical DX sections from the old Antigravity AI Kit README: - Safety Guarantees table (which project files are never touched) - init --force safety features (backup, atomic copy, symlink guard) - update preserved files list (session data, ADRs, rules, contexts) - Agents table grouped by category (23 agents across 6 categories) - Operating Constraints table (5 immutable principles) - Repository structure tree - Author section - Enhanced Quick Start with both options (create new + add to existing) - Getting Started docs page updated with same safety details * fix: address gemini-code-assist review findings on PR #3 H-1: Add JSDoc scope doc + array support to TOML serializer — zero-dep design decision documented, not incomplete by accident. M-1: Extract generateIdeConfigsForInit() helper from initCommand — reduces function length and improves readability. M-2: Replace hardcoded protocol names with generic regex pattern — auto-discovers all "### A. ..." sections from rules.md. M-3: Use finally block in atomicWriteFile for guaranteed temp cleanup — prevents orphaned .tmp files on non-transient errors. M-4: Use exact keyword matching (===) instead of includes() in clusterPatterns — prevents false positives (e.g., "test" matching "contest"). * fix: correct LinkedIn URL in Author section * fix: address gemini-code-assist review #3988546877 findings M-1: Fail explicitly on unknown --ide value instead of silently generating all configs. Lists supported IDEs in error message. M-2: Log error message in create-kit-app IDE config catch block instead of silently swallowing errors. M-3: Log error message in updater IDE config catch block for better debugging and maintainability.
1 participant
Summary
antigravity-ai-kitto@devran-ai/kit(package, CLI, scaffolder, 127 files)Breaking Changes
antigravity-ai-kit@devran-ai/kitag-kitkitcreate-antigravity-appcreate-kit-appToken Optimization Results
Key changes:
workflow-standards.md(deduplicated governance from 22 files)Test plan
npm test— 348/348 tests passingnode bin/kit.js --help— CLI works with new namenode bin/kit.js --version— outputs 4.0.0