davila7/claude-code-templates
41.9
Weak · 16 November 2026
176.3k
lines of production code
Python
with JavaScript, TypeScript
5
measurements over time
What this system is
Features
Add Cloudflare Workers sandbox implementation and tooling
Introduces a new Cloudflare-based sandbox for executing AI-generated code, including the core Worker (\src/index.ts\) that orchestrates Claude AI code generation and isolated execution via the Cloudflare Sandbox SDK. Adds CLI utilities (\launcher.ts\, \monitor.ts\) for running prompts and tracking performance metrics, along with comprehensive documentation (\README.md\, \QUICKSTART.md\, \SANDBOX\_DEBUGGING.md\) and configuration files (\wrangler.toml\, \tsconfig.json\, \.gitignore\, \.dev.vars.example\) to support local development and production deployment.
cli-tool/.claude, cli-tool/.claude/sandbox/cloudflare · high confidence
Add FootballBin match prediction skill for Premier League and Champions League
Users can now access AI-powered match predictions for the Premier League and Champions League through a new skill located in cli-tool/components/skills/sports/footballbin-predictions. This addition includes a SKILL.md definition and a shell script (footballbin.sh) that queries the FootballBin MCP API to retrieve data such as half-time and full-time scores, next goal scorers, corner counts, and key player insights. The skill supports filtering by league, specific matchweeks, and home/away teams, requiring no API keys or environment variables.
cli-tool/components/skills/sports · high confidence
Add Google Analytics analysis Skill
A new Google Analytics skill has been added to the CLI tool, enabling users to analyze website performance, review traffic patterns, and receive data-driven improvement recommendations. This feature includes Python scripts to connect to the Google Analytics 4 Data API using service account credentials, fetch metrics such as sessions, engagement, and conversions, and perform analyses like period comparisons, traffic source breakdowns, and conversion funnel tracking. Users can now ask the tool to review performance over specific timeframes, identify high-bounce pages, and compare current results against previous periods.
cli-tool/components/skills/analytics · high confidence
Add pi-agent-for-claude integration mod
A new mod, \pi-agent-for-claude\, is available that integrates the \pi\ CLI as a native Claude Code subagent. This allows users to delegate tasks to \pi\ directly within the Claude Code agent UI, supporting live streaming, follow-ups, and teammate panes. The integration replaces model requests with detached \pi\ runs, enabling the use of any model \pi\ can reach (such as OpenRouter or Moonshot) while maintaining \pi\'s own tool execution capabilities. Users can install this mod via \npx claude-code-templates@latest --mod integrations/pi-agent-for-claude\ and configure it with specific environment flags for agent teams.
cli-tool/components/mods/integrations/pi-agent-for-claude · high confidence
Add public shared collection pages
Users can now view public collections via a new /c/{slug} route. This page displays the collection name, author, and component count, groups items by type (agents, skills, commands, etc.), and provides a one-click install command for all components in the collection.
dashboard/src/pages/c · high confidence
Added BasePage helper for Playwright Java tests
A new BasePage class has been introduced in the Playwright Java skill components to serve as a foundation for Page Object classes. This class provides common action helpers with Allure step reporting, enforces strict locator strategies (prioritizing role, label, and test ID), and handles navigation and waiting logic, allowing test authors to write more maintainable and consistent UI tests.
repository · high confidence
Added generated todo-app example and Playwright automation utilities
The loki-mode skill now includes a complete, generated full-stack todo application example, featuring an Express backend with SQLite persistence and a React frontend with CRUD operations, alongside a new Playwright skill library providing reusable browser automation helpers for tasks like safe clicking and text input.
cli-tool/components/skills/ai-research/loki-mode/examples/todo-app-generated/backend, cli-tool/components/skills/ai-research/loki-mode/examples/todo-app-generated/frontend, cli-tool/components/skills/utilities/playwright-skill · high confidence
Added static assets for favicon and branding
The dashboard now includes static files for the site favicon, including a web manifest defining Android Chrome icons and a new SVG logo depicting a terminal interface with command prompt elements.
dashboard/public/static · high confidence
Dashboard adds edge caching, security headers, and job data source
The dashboard now serves its static JSON data (components, counts, search index, plugins, trending data, and jobs) with explicit 24-hour edge caching and stale-while-revalidate policies, while Astro assets are cached immutably. Security headers (HSTS, X-Frame-Options, Referrer-Policy, Permissions-Policy) are applied to all routes. A new claude-jobs.json file exposes 85 jobs from Anthropic, HackerNews, RemoteOK, and WeWorkRemotely. Old function-hooks URLs are redirected to the new mods paths.
dashboard/public · high confidence
Introduce /aitmpl side pane for browsing the aitmpl.com catalog
A new productivity mod adds the \/aitmpl\ command, which opens a VS Code Extensions-style side pane for browsing the aitmpl.com component catalog. The pane features a search box, a Browse menu with counts and distinct icons/colors per type (Agents, Skills, Commands, etc.), an Installed section showing locally found agents, commands, skills, and mods, and a Popular/Results list with download counts. Users can view component details, open pages in a browser, or install components via a validated CLI command. The catalog is fetched live and cached for the session, and browsing costs no tokens.
cli-tool/components/mods/productivity/aitmpl · high confidence
Introduce Claude Code Docs Monitor Cloudflare Worker with Telegram alerts and Sentry tracking
A new Cloudflare Worker (\docs-monitor\) is added to automatically monitor the Claude Code documentation page for changes and send notifications to a configured Telegram chat. The worker uses Cloudflare KV to store content hashes and detects updates via SHA-256 comparison, running on a configurable cron schedule (defaulting to every hour). It exposes HTTP endpoints for checking the monitor's status and manually triggering a check. Additionally, the worker includes optional Sentry error tracking and cron health monitoring via a custom \sentry.js\ module, which reports failures and check-in statuses to the 'aitmpl-workers' Sentry project.
cloudflare-workers/docs-monitor · high confidence
Introduce Cloudflare Workers sandbox for AI code generation
Adds a new Cloudflare-based sandbox component that executes AI-generated code (Python and JavaScript) in isolated containers at the edge. This includes a Cloudflare Worker (\src/index.ts\) with \/execute\ and \/health\ endpoints, a CLI launcher (\launcher.ts\) for running prompts, and a real-time monitor (\monitor.ts\) for debugging. The feature requires an Anthropic API key and leverages Cloudflare Durable Objects for sandbox state, with configuration managed via \wrangler.toml\ and environment variables.
cli-tool/components/sandbox/cloudflare · high confidence
Introduce E2B cloud sandbox for isolated Claude Code execution
Users can now execute prompts in an isolated cloud sandbox using E2B, ensuring safe code generation without affecting the local system. This feature requires E2B and Anthropic API keys, which can be provided via CLI arguments or environment variables. The system automatically installs specified components, runs the prompt in a pre-configured Ubuntu environment with Claude Code, and downloads generated files to a local \e2b-output\ directory. It includes extended timeouts (up to 15 minutes) and retry logic for WebSocket connection stability, with detailed logging and a real-time monitor for debugging.
cli-tool/components/sandbox/e2b · high confidence
Introduce create-claude-config CLI with extensive dashboard and component management options
A new CLI entry point, create-claude-config, is added to manage Claude Code configurations and global AI agents. It provides a wide range of dashboard launchers for analytics, mobile chats, agent teams, plugins, skills, and a 2025 year-in-review summary, alongside options to run health checks and enable cloud tunneling. The tool also supports granular installation of specific components (agents, commands, MCPs, settings, hooks, skills, loops, and mods) and allows executing prompts or cloning shared sessions directly from the command line.
cli-tool/bin · high confidence
Introduce jev-model-router mod for intelligent model and effort routing
The new jev-model-router mod automatically selects the appropriate model and reasoning effort for each task by leveraging TypeSafe's Jev decision model or the Vercel AI Gateway. It classifies prompts at submission to route subagent models and adjust the main loop's effort and model based on task complexity (fast, balanced, or deep) and risk, using calibrated confidence thresholds to prevent costly misrouting. The mod supports configuration for API keys, backend selection, tier-specific models, and confidence bars, with a fail-open design that falls back to the engine's built-in classifier if no external keys are set.
cli-tool/components/mods/productivity/jev-model-router · high confidence
Introduce jev-skill-suggestion mod to reduce context usage and improve skill selection
The new jev-skill-suggestion mod intercepts the engine's skill listing to hide it from the model, significantly reducing context window usage. Instead of the full listing, the mod uses TypeSafe's Jev decision model (or a fallback built-in classifier) to rank available skills and suggest at most one relevant skill per prompt. The selected skill's instructions are then injected directly into the prompt, allowing users to hide skills from the model's automatic loading while keeping them available for manual invocation via the \/name\ command.
cli-tool/components/mods/productivity/jev-skill-suggestion · high confidence
Introduce ultra-config: five settings profiles with a side-pane UI
A new productivity mod adds a side pane (accessible via /ultra-config) offering five pre-configured settings profiles—Token saver, Fast flow, Max quality, Guarded, and Autopilot. Each profile defines a set of changes to Claude Code's /config rows (like model, effort, output style) and direct settings.json keys (like subagent model, environment caps, and permission deny rules). Users can preview exactly what each profile will change, activate it with one press, and undo the last activation. The mod writes /config rows via the engine's config API and persists other settings to \~/.claude/settings.json (with a backup copy), ensuring changes are applied correctly even when the engine reloads settings asynchronously. It requires Claude Code 2.1.287 or later.
cli-tool/components/mods/productivity/ultra-config · high confidence
Introduces Obsidian vault automation scripts for linking, tagging, and metadata
Adds a suite of Python scripts to the Obsidian Ops Team agent to automate vault maintenance. The new tools include a Link Suggester and keyword/entity connection finders to identify and insert bidirectional links between notes, an Enhanced Tag Standardizer to consolidate and hierarchically organize tags, a Daily Notes Connector to create temporal and content-based connections, a Metadata Adder to inject standardized frontmatter, an MOC Generator to create Map of Content files, and utility scripts to fix quoted tags and implement entity-based connections from reports.
cli-tool/components/agents/obsidian-ops-team · high confidence
Introducing jev-skill-typeahead: live skill and subagent suggestions while typing
A new productivity mod, jev-skill-typeahead, displays a band above the prompt box that ranks skills and subagents Claude may call as you type. It uses instant keyword matching (supporting English and Spanish) and, after a configurable pause, sends the draft to a decision backend (TypeSafe's Jev, the Vercel AI Gateway, or Claude Code's built-in classifier) to mark the most likely candidate. The mod includes a /jev-skill-typeahead command to open a details pane with full names, descriptions, and matched terms, and can optionally attach the selected skill to the model upon submission. It is available for Claude Code 2.1.287+.
cli-tool/components/mods/productivity/jev-skill-typeahead · high confidence
Modernized analytics dashboard with real-time updates and theme support
The analytics web interface has been rebuilt using Astro, introducing a new dashboard layout with a dark/light theme toggle and real-time data synchronization via WebSocket. Users now see live updates for conversations, token usage, and agent activity, with a dedicated Agent Analytics view and an interactive 2025 Year-in-Review visualization. Also in this area: 1 commit (1 fix) fixing cli-tool, cli-tool/analytics-ui.
cli-tool/src/analytics-web · medium confidence · unverified
New 2048 and cc-arcade game mods available
Users can now play the 2048 puzzle game and a collection of nine arcade-style games (including Snake, Tetris, Doom, and Minesweeper) directly above the Claude Code prompt. These mods, sourced from the cc-arcade project, run as function hooks that respond to keyboard and mouse input without consuming tokens. The games automatically pause when Claude finishes a turn and resume when the user interacts, while also persisting best scores across sessions.
cli-tool/components/mods/games · high confidence
New AI research skills for agent evaluation, management, and orchestration frameworks
Added documentation and setup guides for several new AI research skills in the CLI tool. This includes agent evaluation and memory systems (agent-evaluation, agent-memory-mcp, agent-memory-systems), an agent manager for tmux-based session control, and an agent tool builder for designing LLM tools. It also introduces comprehensive guides for major agent orchestration platforms: AutoGPT (visual workflow builder, custom blocks, execution patterns), CrewAI (multi-agent collaboration, flows, tools, troubleshooting), and LangChain (agents, RAG, vector store integrations, observability).
cli-tool/components/skills/ai-research · high confidence
New API endpoints for managing and sharing collections
This change introduces a new set of Astro-based API routes in the dashboard to handle collection lifecycle and sharing. Users can now create, read, update, and delete their own collections via the index and \[id\] endpoints, which enforce authentication and name length limits. A new items endpoint allows adding, removing, and moving components between collections. Additionally, a share endpoint enables users to toggle public visibility on collections, generating unique slugs for access. The public GET share endpoint allows unauthenticated users to view collection details and items by slug, exposing only safe data like component paths and the author's display name.
dashboard/src/pages/api/collections · high confidence
New Astro-based analytics dashboard with real-time updates and mobile chat interface
The CLI tool now includes a modernized analytics dashboard built with Astro and React, replacing the previous implementation. This new web interface provides real-time usage monitoring via WebSocket connections, featuring an activity heatmap, token usage breakdowns, session tracking, and agent statistics. The dashboard supports dark/light theme toggling, time-range filtering (7/30/90 days), and includes a dedicated mobile-optimized chat interface for reviewing conversation history. Static assets are generated to the src/analytics-web directory, and the UI includes comprehensive loading states and skeleton screens for improved perceived performance.
cli-tool/src · high confidence
New Cloudflare Worker for scheduled health and code checks with Sentry monitoring
A new Cloudflare Worker named 'aitmpl-crons' has been introduced to replace previous Vercel-based cron jobs. This worker executes two scheduled tasks: a 'claude-code-check' every 30 minutes and a 'health-check' every hour, both calling specific dashboard API endpoints. The implementation includes built-in error tracking via Sentry, sending detailed reports and cron monitor check-ins to alert on failures or missed schedules, while also providing a manual trigger endpoint for testing.
cloudflare-workers/crons · high confidence
New DOCX and documentation skills added to document-processing toolkit
The document-processing area now includes two new skills: a 'doc' skill for reading, creating, and editing .docx files with visual layout validation (using python-docx and LibreOffice), and a 'docx-official' skill for professional Word document creation and editing with advanced features like tracked changes, comments, and styling. Additionally, a 'documentation-templates' skill provides structured templates for READMEs, API docs, changelogs, and architecture decision records. These skills come with supporting scripts for DOCX rendering, tracked change acceptance, comment insertion, and XML manipulation, along with appropriate licensing (Apache 2.0 for doc, proprietary for docx-official).
cli-tool/components/skills/document-processing · high confidence
New Docker-based sandbox for isolated Claude Code execution
Users can now run Claude Code in an isolated Docker container via the \cli-tool/components/sandbox/docker\ module. This feature introduces a launcher (\docker-launcher.js\) that builds a Node.js 22 Alpine image (using \@anthropic-ai/claude-agent-sdk\) and executes prompts inside it, while an executor script (\execute.js\) handles component installation and file capture. Generated output is automatically copied to a local \output/\ directory, allowing users to iterate on code generation in a secure, non-root environment.
cli-tool/components/sandbox/docker · high confidence
New DoorDash CLI skills for dietary safety, group ordering, and spending controls
This change introduces a suite of new skills for the DoorDash CLI (dd-cli) located in cli-tool/components/skills/doordash. The doordash-allergy-shield skill adds a persistent dietary profile and mandatory cart vetting to prevent allergen exposure, using a hook to double-check high-severity items at checkout. The doordash-group-orders skill enables coordinated team ordering by maintaining a roster, attributing cart items to specific people, and calculating cost splits. The doordash-spend-guard skill enforces deterministic spending limits (per-order, daily, weekly, monthly) via a wrapper script (dd-guard.sh) that blocks out-of-policy checkouts. Additional skills include doordash-order-playbooks for saving and safely recalling orders with drift detection, and doordash-order-ledger for auditing agent activity and spending history.
cli-tool/components/skills/doordash · high confidence
New Linear tickets and Neon database branch mods
Two new integration mods are now available. The linear-tickets mod displays your Linear tickets in a side pane with a status bar, a closed-per-day chart, and a cycle burn-up view; it is read-only, requires a Linear API key, and is opened via the /linear command. The neon-branch-per-session mod creates a dedicated Neon database branch for each Claude Code session, sets DATABASE\_URL for all commands, and manages the branch's lifecycle (create, keep, delete) via the /neon command.
cli-tool/components/mods/integrations/linear-tickets · high confidence
New Mods for search, observability, and package management
Three new Mods are available to extend Claude Code's capabilities. The websearch-to-exa Mod overrides the built-in WebSearch tool to route queries to the Exa search API, with automatic fallback to the default search if no API key is configured or if Exa fails. The universal-audit-log Mod records every engine and plugin event (tool calls, prompts, file system, HTTP, and process calls) as JSON lines in a configurable file, including timing and outcome details. The npm-to-pnpm-rewriter Mod automatically rewrites npm and npx commands to pnpm (or yarn/bun) before execution, allowing the model to use standard npm syntax while the project uses a different package manager.
(repo-wide) · high confidence
New OWASP Security Skill for comprehensive security guidance
A new 'Comprehensive OWASP Security Skill' has been added to the .claude-plugin/skills/owasp-security directory. This skill provides developers with unified security guidance across six OWASP standards: Top 10 (2021) for web apps, ASVS 5.0, MASVS v2.1.0 for mobile, API Security Top 10 (2023), Kubernetes Top 10 (2022), and Agentic Applications 2026 for AI/LLM systems. The skill includes a detailed reference guide (SKILL.md) and nine example files demonstrating vulnerable versus secure implementations for issues like broken access control, cryptographic failures, injection, security misconfiguration, and prompt injection. Users can now trigger deep security analysis for code reviews, architecture decisions, and deployment configurations by dropping this skill into their assistant's skill directory.
.claude-plugin · high confidence
New Plugins & Marketplaces dashboard pages
The dashboard now includes dedicated pages for browsing and installing Claude Code plugins. The main listing at /plugins displays a grid of available plugin marketplaces and collections, while individual detail pages at /plugins/\[slug\] provide specific installation instructions, GitHub links, and metadata for each marketplace. These pages are statically generated at build time by reading from public/plugins.json.
dashboard/src/pages/plugins · high confidence
New Python-based skills added to the CLI tool
This update adds several new Python scripts to the CLI tool's skills library, expanding capabilities in AI research, development, security, document processing, and enterprise communication. Specifically, it introduces a prompt optimization script for A/B testing and iterative improvement, a production-ready FastAPI REST API template with pagination and security middleware, and security compliance tools for risk assessment (including ALE calculations and cost-benefit analysis) and vulnerability prioritization based on CVSS and business context. Additionally, it adds a markdown-based task manager for tracking implementation phases, bounding box validation scripts for PDF form fields, a PowerPoint text inventory extractor with formatting preservation, and a script to generate daily meeting digests from Claude Code session logs.
python · high confidence
New TypeScript type definitions for the Claude Code plugin API
A new \claude-code.d.ts\ file has been added to provide TypeScript declarations for the plugin API, enabling developers to write type-safe hooks modules. This file defines the core interfaces for the engine, including \AgentCallRecord\, \AgentInfo\, and \AgentLoop\, and documents the testing environment and module structure required for building mods. Also in this area: 11 commits (1 fix) modifying cli-tool.
cli-tool/components/mods/types · medium confidence · unverified
New accessibility and agent-development skill guides
Added two new skill definitions under the development category: an accessibility skill that provides WCAG 2.1 audit guidelines and implementation patterns for web accessibility, and an agent-development skill that includes documentation, examples, and a validation script for creating and managing autonomous agent configurations.
cli-tool/components/skills/development · high confidence
New automation and deployment scripts for the dashboard and content generation
The repository now includes a suite of new scripts in the \scripts/\ directory to support the Astro dashboard and content workflows. \deploy.sh\ provides a CLI for deploying the unified Astro dashboard to Vercel, requiring \VERCEL\_ORG\_ID\ and \VERCEL\_DASHBOARD\_PROJECT\_ID\ environment variables. \dev-server.js\ offers a local Express-based development server for the \docs/\ directory, serving static files and handling routing for agents, commands, hooks, and blog posts. Content generation is automated via \generate\_agents\_api.py\ (creating a lightweight agents API from \components.json\), \generate\_plugins\_json.py\ (scanning GitHub for \.claude-plugin\ data to build the marketplace), \generate\_components\_json.py\ (which now includes security validation via \npm run security-audit:json\), \generate\_trending\_data.py\ (fetching download stats from Supabase with retry logic and type aliasing for legacy hooks), \generate\_claude\_jobs.py\ (scraping free job sources for Claude Code positions), \generate\_claude\_prs.py\ (tracking PRs from \claude/\ branches), \generate\_star\_history.py\ (rendering a self-hosted, theme-aware SVG stargazer chart), and \generate\_blog\_images.py\ / \generate\_blog\_images\_v2.py\ (using Google Gemini Imagen API to generate blog cover images and diagrams). Also in this area: 4 commits (0 fixes) modifying scripts; 2 commits (1 fix) modifying scripts; Improved component metadata parsing and added validation script.
scripts · medium confidence · unverified
New automation hooks for Claude Code: agent config, build triggers, logging, and notifications
This update adds a suite of new automation hooks to the CLI tool. Users can now automatically load project-specific agent behavior via an AGENTS.md file at session start. The tool can also trigger builds (npm, yarn, make, cargo, maven, gradle) after file edits, log all file mutations to a CSV for review, and check dependencies for security vulnerabilities when package files are modified. Additionally, it introduces comprehensive notification hooks for Discord, Slack, and Telegram, offering both simple finish notifications and detailed session summaries (including duration and memory usage), as well as alerts for long-running operations. A new 'Plan Mode Game' hook allows users to play browser-based games (Dino Runner, Snake, Flappy Bird) with a live activity panel while Claude is in planning mode.
cli-tool/components/hooks/automation · high confidence
New business-marketing skills for app building, AI product development, and agile workflows
This update adds a suite of new skills to the business-marketing category, providing specialized guidance for application development, AI integration, and product management. The App Builder skill introduces an orchestrator for full-stack applications, including project detection logic, agent coordination pipelines, scaffolding structures, and templates for stacks like Next.js, Flutter, and Electron. AI-focused skills (ai-product, ai-wrapper-product) offer patterns for LLM integration, prompt engineering, and cost management. Additional skills cover A/B test setup, analytics tracking, and agile product ownership, including a Python-based user story generator script.
cli-tool/components/skills/business-marketing · high confidence
New career skills category with 20 job-search workflow tools
Users now have access to a new 'career' skills category containing 20 specialized tools for the job-search workflow. This update adds capabilities for creating and formatting academic CVs, translating skills for career changers, generating personalized cover letters, designing ATS-compatible creative resumes, writing executive-level resumes, preparing for interviews with STAR stories, analyzing job descriptions for fit, optimizing LinkedIn profiles, comparing multiple job offers, and writing portfolio case studies.
cli-tool/components/skills/career · high confidence
New creative design skills added: 3D web, accessibility, algorithmic art, and C4 architecture
The CLI tool now includes four new skills in the creative-design category. The 3D Web Experience skill provides guidance for building interactive 3D scenes using Three.js, React Three Fiber, and Spline. The Accessibility Auditor skill offers expert guidance for WCAG 2.1 compliance, ARIA implementation, and inclusive design. The Algorithmic Art skill enables the creation of generative art using p5.js with seeded randomness and interactive parameter exploration. The C4 Architecture skill generates software architecture documentation using C4 model diagrams in Mermaid syntax.
cli-tool/components/skills/creative-design · high confidence
New daily health report worker sends Telegram digest with Sentry noise filtering
A new Cloudflare Worker at cloudflare-workers/daily-health-report has been added to provide a proactive daily monitoring heartbeat. Scheduled to run daily at 14:00 UTC, it aggregates dashboard site health (via GET /api/health-check) and a 24-hour Sentry error summary across three projects (aitmpl-workers, aitmpl-dashboard, aitmpl-cli) into a single Telegram digest. A key behavioral feature is the automatic resolution of known test noise: issues with titles matching 'manual verification test' or 'manual verification' are silently marked as resolved in Sentry before reporting, ensuring the digest only highlights genuine production errors. The worker also exposes a POST /trigger endpoint for manual execution and a GET /status endpoint for health checks.
cloudflare-workers/daily-health-report · high confidence
New design-to-code skill for converting Figma designs to React
A new 'design-to-code' skill has been added to the CLI tool, enabling the conversion of Figma designs into production-ready React, TypeScript, Vite, and TailwindCSS components. This feature introduces a helper script (\coderio-skill.mjs\) and a structured workflow (Setup, Protocol Generation, Code Generation) that automates fetching Figma data, generating component structures and props, and producing code with high visual fidelity. Users can now leverage this automated pipeline to restore UI designs with robust error handling and checkpoint recovery.
cli-tool/components/skills/design-to-code · high confidence
New development team agent roles added
The CLI tool now includes a set of specialized agent definitions within the \cli-tool/components/agents/development-team\ directory to guide AI-assisted development. New roles include \backend-architect\ for system design and API strategy, \backend-developer\ for server-side implementation, \cli-ui-designer\ for terminal-style interfaces, \code-architect\ for feature blueprints, \code-explorer\ for codebase analysis, \devops-engineer\ for CI/CD and infrastructure, \electron-pro\ for desktop applications, \flutter-ui-developer\ for cross-platform mobile/web UIs, and an updated \frontend-developer\ covering modern React, Vue, and Angular stacks. These agents provide structured workflows, best practices, and output templates to standardize how the CLI tool handles various development tasks.
cli-tool/components/agents/development-team · high confidence
New development tool hooks for debugging, auditing, and workflow automation
This update adds a suite of new hook components to the development-tools directory, enabling several new capabilities for Claude Code users. The debug-window hook now provides an auto-managed terminal window that tails the debug log on session start (when the --debug flag is used) and closes it on session end, with an option to disable auto-closing. New audit and tracking hooks include change-tracker (logs file modifications/creations to a global log), command-logger (logs all tool usage to a global log), and edit-audit-log (logs edits to a project-local file). Workflow automation is enhanced with file-backup (creates timestamped backups before edits), lint-on-save (runs ESLint, Pylint, or RuboCop after edits), smart-formatting (runs Prettier, Black, gofmt, etc. after edits), and nextjs-code-quality-enforcer (validates Next.js best practices). Additionally, a worktree-ghostty hook is introduced for macOS, which automates git worktree creation and opens a 3-panel Ghostty layout with lazygit and yazi.
cli-tool/components/hooks/development-tools · high confidence
New dynamic featured integration detail pages
The dashboard now includes a new \\[slug\].astro\ page that renders detailed, dynamic views for featured integrations (such as Bright Data, ClaudeKit, and Braingrid). This page displays a redesigned layout with a verified badge, social proof indicators, and a prominent call-to-action, pulling content from the \FEATURED\_ITEMS\ constant to provide users with a structured overview and setup guide for each integration.
dashboard/src/pages/featured · high confidence
New enterprise communication skills for API handoff, standups, and compliance
Added several new skills to the enterprise-communication category: a backend-to-frontend handoff documentation generator that creates structured API integration docs from completed backend code; a daily meeting update tool that conducts an interactive standup interview and optionally pulls activity from GitHub, Jira, and Claude Code session history; a CAPA (Corrective and Preventive Action) officer skill for quality management system investigations and root cause analysis; a brand guidelines skill applying Anthropic's official colors and typography to artifacts; and a data privacy compliance skill covering GDPR, CCPA, and HIPAA requirements.
cli-tool/components/skills/enterprise-communication · high confidence
New enterprise security and productivity mods introduced
Two new mods are now available for Claude Code 2.1.287+. The admin-capability-lockdown mod provides organization-level control by withholding network and process access from plugins, enforcing plugin allowlists, and restricting the Bash tool (with configurable deny or guardrail modes). The webfetch-cache mod improves productivity by caching WebFetch results within a session, including support for paginated reads via offset, with configurable TTL and cache size limits.
cli-tool/components/mods/enterprise · high confidence
New media skills for image enhancement, screenshot capture, and speech generation
The CLI tool now includes three new media skills in the \cli-tool/components/skills/media\ directory. The \image-enhancer\ skill improves image quality by upscaling resolution, sharpening details, and reducing compression artifacts. The \screenshot\ skill provides cross-platform desktop capture capabilities (macOS, Linux, Windows) with support for full-screen, window-specific, and region-based captures, including a macOS preflight helper to manage Screen Recording permissions. The \speech\ skill enables text-to-speech generation using OpenAI's \gpt-4o-mini-tts-2025-12-15\ as the default provider, with optional support for Atlas Cloud TTS when explicitly requested.
cli-tool/components/skills/media · high confidence
New monitoring hooks for context visualization, desktop notifications, and LangSmith tracing
Added three new hook configurations to the monitoring directory: a context-timeline visualizer that runs a local HTTP server to display real-time context window usage and tool execution as a git-graph timeline; a desktop notification hook that sends native OS notifications (via osascript or notify-send) when a session stops; and a LangSmith tracing integration that exports conversation traces to LangSmith for monitoring and analysis.
cli-tool/components/hooks/monitoring · high confidence
New productivity skills added to the CLI toolkit
The CLI tool now includes several new productivity skills in the \cli-tool/components/skills/productivity\ directory. These include \academy-guide\ for recommending learning resources from Claude Academy, \avoid-ai-writing\ for auditing and rewriting content to remove AI writing patterns, \address-github-comments\ for systematically handling PR feedback via the GitHub CLI, \brainstorming\ for collaborative design exploration, \claude-code-sessions\ for searching and managing session history, \code-review-excellence\ for structured code review workflows, \code-simplifier\ for refactoring code for clarity, and \commit-work\ for creating high-quality git commits.
cli-tool/components/skills/productivity · high confidence
New programmatic video creation skills for Remotion, Motion Canvas, and Manim
The CLI tool now includes dedicated skill sets for three programmatic video creation frameworks. The Remotion skill provides guidance for building videos using React, covering animations, media assets, captions, and 3D integration. The Motion Canvas skill offers a production-ready setup for TypeScript-based animations, including specific workarounds for ESM/CommonJS interoperability. The Manim skill supports the creation of mathematical and educational animations using Python. These skills provide the necessary documentation and best practices for generating video content programmatically within the tool.
cli-tool/components/skills/video · high confidence
New prompt-cache-control observability mod
A new mod adds a prompt-cache meter above the Claude Code prompt, displaying per-request token usage (read, write, uncached), a live countdown to cache expiry, and actionable advice (keep going, /compact, or /clear). The mod follows Claude Code's TTL rules (1 hour for subscriptions, 5 minutes otherwise) and infers the actual lifetime from request timing. It includes a /cache pane with a per-turn table, configurable warnings, and idle compaction countdowns for Claude Code 2.1.290+.
cli-tool/components/mods/observability/prompt-cache-control · high confidence
New security assessment skills added to the CLI tool
The CLI tool now includes a new set of security-focused skills in the \cli-tool/components/skills/security\ directory. These new capabilities allow users to perform specialized security assessments, including Active Directory attack simulations, API fuzzing and bug bounty hunting, AWS and general cloud penetration testing, and broken authentication testing. The update also adds skills for API security best practices, API security testing workflows, and Burp Suite web application testing, providing structured guidance and command examples for red team and security audit operations.
cli-tool/components/skills/security · high confidence
New security hooks for command, file, and secret protection
Added a suite of security hooks in the CLI tool to prevent accidental data loss and credential exposure. The dangerous-command-blocker and shell-wrapper-guard hooks intercept Bash commands to block catastrophic operations (like recursive deletes or disk formatting) and detect destructive commands hidden inside interpreter wrappers (e.g., python3 -c, sh -c). The secret-scanner hook prevents commits containing hardcoded API keys, tokens, and passwords from providers like AWS, GitHub, and Stripe. Additionally, file-protection and env-file-protection hooks block modifications to critical system paths and .env files, while the force-push-blocker prevents accidental git history overwrites. A post-edit security-scanner hook runs static analysis tools (semgrep, bandit, gitleaks) if available to catch vulnerabilities in modified files.
cli-tool/components/hooks/security · high confidence
New security mods: block-destructive-commands and jev-auto-mode
Two new security mods are available for Claude Code (requires version 2.1.259+ with function hooks enabled). The \block-destructive-commands\ mod intercepts Bash tool calls to deny dangerous patterns—such as recursive root deletion, force pushes, and disk formatting—before they execute, with support for custom regex patterns. The \jev-auto-mode\ mod introduces a comprehensive JSON-policy permission layer that governs all tool calls, slash commands, and skills. It enforces rules (deny/ask/allow) and, for unmatched actions, delegates to TypeSafe's Jev judge (or a built-in classifier) to assess hazards like exfiltration or scope violations, allowing users to define granular policies via user and project-level JSON files. Also in this area: Mods are now generally available and the api-architect agent is updated.
cli-tool/components/mods/security/jev-guardrails, cli-tool/components/mods/security/jev-vercel-sandbox · high confidence
New security validation system for components
The CLI tool now includes a comprehensive security validation system for agents, commands, MCPs, settings, and hooks. This new capability introduces a five-tier validation framework: Structural (YAML frontmatter, file size, encoding), Integrity (SHA256 hashing and tamper detection), Semantic (prompt injection, jailbreak, and credential harvesting detection), Reference (URL protocol validation and private IP blocking), and Provenance (author and repository metadata verification). The system is orchestrated by ValidationOrchestrator and accessible via the new \--security-audit\ CLI command, which supports verbose output, JSON reporting, and CI/CD integration modes.
cli-tool/src/validation · high confidence
New session-time-machine mod for forking sessions from a timeline
This change introduces the \session-time-machine\ mod, which adds a \/timemachine\ command to visualize the current session as a timeline of prompts, tool calls, and turn ends. Users can select any point in the timeline to fork the session: the mod snapshots the working tree using git worktrees and creates a new session with the transcript cut at that point, providing a \claude --resume\ command to continue the forked session without affecting the original. The mod includes a UI pane with color-coded rows, configuration options for copying commands and opening terminals, and hooks into the session lifecycle to manage snapshots and transcript parsing.
cli-tool/components/mods/productivity/session-time-machine · high confidence
New statusline components for Vercel, Cloudflare, Neon, and game dev workflows
The CLI tool now includes a large set of new statusline configurations that provide real-time monitoring and gamified feedback directly in the editor. New integrations include a Vercel deployment monitor with clickable deploy links, a Cloudflare Pages deployment monitor, and Neon database context/development/resource monitors that display branch, compute state, and usage costs. Game development workflows are supported with statuslines for Unity, Unreal, and Godot project dashboards, asset pipeline controllers, and multi-platform build trackers. Additionally, several playful, session-based statuslines have been added, including a coding fitness tracker, RPG-style level progression, a virtual pet (tamagotchi), a code casino, and a bug circus, alongside practical utilities like a context usage monitor and a deadline countdown timer.
cli-tool/components/settings/statusline · high confidence
New subagent-tuner mod for controlling subagent model and effort
A new productivity mod, subagent-tuner, has been added to allow users to control the model and reasoning effort of every subagent Claude Code starts. It offers two modes: Manual, which presents a recommendation and allows the user to override or save a default via a pane or dialog, and Auto, which automatically applies a saved default or a recommendation based on the task kind (e.g., review, build, search) without asking. The mod includes a UI band for quick settings changes and persists user preferences.
cli-tool/components/mods/productivity/subagent-tuner · high confidence
New utility skills added for browser automation, web crawling, and development workflows
The CLI tool now includes a set of new utility skills in the \cli-tool/components/skills/utilities\ directory. These include \browser-automation\ and \playwright-skill\ for web testing and scraping using Playwright, \cf-crawl\ for crawling websites via the Cloudflare Browser Rendering API (with support for incremental crawls via \modifiedSince\), \browser-extension-builder\ for creating Manifest V3 extensions, \busybox-on-windows\ for running UNIX tools on Windows, \domain-name-brainstormer\ for generating and checking domain availability, \geo-fundamentals\ for optimizing content for AI search engines (including a \geo\_checker.py\ audit script), and \network-101\ for configuring network services like HTTP, HTTPS, SNMP, and SMB in penetration testing labs.
cli-tool/components/skills/utilities · high confidence
New web-development skills for Astro, blog generation, and external AI routing
Added new skill definitions to the CLI tool's web-development library: an Astro framework skill covering islands architecture and content collections; a comprehensive 'building-blog' skill that orchestrates adding a Next.js + Sanity blog with technical requirements and image style guides; a 'blockrun' skill enabling external AI capabilities (image generation, X/Twitter data) via micropayments; and skills for Chrome Extension (Manifest V3), Drizzle ORM, Electron, and Exa Search. These additions expand the tool's ability to guide users through specific modern web development stacks and external service integrations.
cli-tool/components/skills/web-development · high confidence
New weekly newsletter worker (paused)
A new Cloudflare Worker (\cloudflare-workers/newsletter\) has been added to compose and send a weekly email featuring trending community components via Resend Broadcasts. The worker selects one item per category (Skill, Agent, MCP, Hook, Setting) using weighted randomness and rotates email copy to avoid repetition. It includes endpoints for manual triggering (\/trigger\) and content preview (\/preview\), along with built-in unsubscribe handling and optional Sentry error tracking. The scheduled weekly send is currently paused (disabled via \crons = \[\]\ in \wrangler.toml\ and \NEWSLETTER\_ENABLED = "false"\) because the emails were being flagged as spam; the worker remains deployed for iteration and can be re-enabled by restoring the cron schedule and setting the environment variable to true.
cloudflare-workers/newsletter · high confidence
Telegram notifications for new pull requests
A new hook script has been added to automatically send a Telegram notification whenever a pull request is created. The notification includes the PR URL, the current branch name, and the Vercel preview URL (if configured via environment variables), allowing users to stay informed about new PRs directly in their chat.
.claude/hooks · high confidence
Behavioural changes
1 commit (0 fixes) modifying cli-tool/components/mods/integrations/rex-tty
A change to existing behaviour in cli-tool/components/mods/integrations/rex-tty — 1 commit, 7 files: “feat(mods): rex-tty, run commands that need a terminal in a Rex pane (\#1517)” (refs \#1517).
cli-tool/components/mods/integrations/rex-tty · medium confidence · unverified
1 commit (0 fixes) modifying cli-tool/components/mods/security/human-gate, cli-tool/components/mods/ui/live-preview
A change to existing behaviour in cli-tool/components/mods/security/human-gate, cli-tool/components/mods/ui/live-preview — 1 commit, 12 files: “feat(mods): add human-gate, policy-ceiling and live-preview (\#1366)” (refs \#1366).
cli-tool/components/mods/productivity/session-chats, cli-tool/components/mods/security/human-gate, cli-tool/components/mods/ui/live-preview, cli-tool/components/mods/ui/spooky-season · medium confidence · unverified
10 commits (3 fixes) modifying dashboard/src/lib/api
A change to existing behaviour in dashboard/src/lib/api — 10 commits (3 fixs), 7 files: “feat(cli): 1.33.0 install fixes (mods from aitmpl.com, short names, --search, system-folder guard) (\#1467)”; “feat(tracking): record whether an AI agent ran each install (rebased \#1270 + endpoint tests) (\#1368)”; “fix(dashboard): read visitor country from cf-ipcountry (\#1253)”; and 7 more (refs \#1467, \#1270, \#1368, \#1253, \#1429, …).
dashboard/src/lib/api · medium confidence · unverified
15 commits (4 fixes) modifying dashboard/src/pages/api
A change to existing behaviour in dashboard/src/pages/api — 15 commits (4 fixs), 7 files: “feat: Function Hooks section (experimental) — 10 hooks, CLI flag, blog (\#867)”; “feat(tracking): record whether an AI agent ran each install (rebased \#1270 + endpoint tests) (\#1368)”; “feat: Add dashboard source code (app.aitmpl.com)”; and 12 more (refs \#867, \#1270, \#1368, \#1446, \#1253, …).
dashboard/src/pages/api · medium confidence · unverified
2 commits (0 fixes) modifying cli-tool/components/hooks/quality-gates
A change to existing behaviour in cli-tool/components/hooks/quality-gates — 2 commits, 6 files: “hooks: tdd-gate: enforce the skill's red-green cycle (\#1415)”; “feat: enterprise quality hooks and agents from pm-workspace (\#397)” (refs \#1415, \#397).
cli-tool/components/hooks/quality-gates · medium confidence · unverified
2 commits (1 fix) modifying .well-known, dashboard/public/.well-known
A change to existing behaviour in .well-known, dashboard/public/.well-known — 2 commits (1 fix), 2 files: “fix(funding): serve funding.json from www.aitmpl.com for FLOSS/Fund (\#1495)”; “chore: add funding.json manifest for FLOSS/fund (\#1359)” (refs \#1495, \#1359).
.well-known, dashboard/public/.well-known · medium confidence · unverified
2 commits (1 fix) modifying cli-tool/components/mods/integrations/bluesky-feed
A change to existing behaviour in cli-tool/components/mods/integrations/bluesky-feed — 2 commits (1 fix), 6 files: “fix(mods): ignore subagent turns at turn.start (\#1449)”; “feat(mods): add bluesky-feed, your Bluesky timeline in a pane (\#1207)” (refs \#1449, \#1207).
cli-tool/components/mods/integrations/bluesky-feed · medium confidence · unverified
2 commits (1 fix) modifying cli-tool/components/mods/ui/turn-inspector
A change to existing behaviour in cli-tool/components/mods/ui/turn-inspector — 2 commits (1 fix), 6 files: “fix(mods): ignore subagent turns at turn.start (\#1449)”; “feat(mods): turn-inspector, a turn-by-turn view of the context window (\#1167)” (refs \#1449, \#1167).
cli-tool/components/mods/ui/turn-inspector · medium confidence · unverified
23 commits (0 fixes) modifying docs-site
A change to existing behaviour in docs-site — 23 commits, 38 files: “Ship the upstream LICENSE/NOTICE with external skills (\#1577)”; “Add Neon postgres-best-practices as an external skill (database) (\#1570)”; “feat(cli): 1.33.0 install fixes (mods from aitmpl.com, short names, --search, system-folder guard) (\#1467)”; and 4 more (refs \#1577, \#1570, \#1467, \#1575, \#1438, …).
docs-site · medium confidence · unverified
23 commits (4 fixes) modifying dashboard/src/layouts
A change to existing behaviour in dashboard/src/layouts — 23 commits (4 fixs), 1 file: “feat(dashboard): WebMCP V1 — expose catalog as read-only agent tools (\#850)”; “feat(dashboard): site-wide share card for non-component pages (\#1344)”; “feat(seo): Add critical SEO foundations — meta tags, canonical, OG, JSON-LD, robots.txt”; and 20 more (refs \#850, \#1344, \#1563, \#1153, \#1399, …).
dashboard/src/layouts · medium confidence · unverified
3 commits (0 fixes) modifying dashboard/src/pages/s
A change to existing behaviour in dashboard/src/pages/s — 3 commits, 1 file: “feat(dashboard): harness flow and layers on /s/ stack pages (\#1502)”; “Stack diagram: tabs, neutral palette, component hover cards, ordered layers (\#1505)”; “feat(dashboard): /s/\<code\> stack links, no account or database (traction initiative 9) (\#1471)” (refs \#1502, \#1505, \#1471).
dashboard/src/pages/s · medium confidence · unverified
32 commits (9 fixes) modifying dashboard/src/pages/component
A change to existing behaviour in dashboard/src/pages/component — 32 commits (9 fixs), 1 file: “feat: Function Hooks section (experimental) — 10 hooks, CLI flag, blog (\#867)”; “feat: external skills installed from their author's repo (first: t3dotgg/tstack) (\#1279)”; “fix(skills): move non-spec frontmatter keys under metadata (299 skills) (\#1258)”; and 29 more (refs \#867, \#1279, \#1258, \#1388, \#1456, …).
dashboard/src/pages/component · medium confidence · unverified
38 commits (5 fixes) modifying dashboard/src/pages
A change to existing behaviour in dashboard/src/pages — 38 commits (5 fixs), 15 files: “feat: Function Hooks section (experimental) — 10 hooks, CLI flag, blog (\#867)”; “feat(dashboard): /security page with an animated skill review flow (\#1518)”; “feat(dashboard): re-tematizar UI a estética CLI/terminal híbrida”; and 35 more (refs \#867, \#1518, \#1337, \#921, \#1388, …).
dashboard/src/pages · medium confidence · unverified
4 commits (0 fixes) modifying cli-tool/components/skills/web-data
A change to existing behaviour in cli-tool/components/skills/web-data — 4 commits, 27 files: “Add the four Gaffa skills as external skills (\#1383)”; “Add PriceWin MCP and skill to web-data (\#1060)”; “feat: add Bright Data web data template as featured integration (\#413)”; and 1 more (refs \#1383, \#1060, \#413, \#760).
cli-tool/components/skills/web-data · medium confidence · unverified
4 commits (1 fix) modifying cli-tool/components/mods/integrations/neon-branch-per-session
A change to existing behaviour in cli-tool/components/mods/integrations/neon-branch-per-session — 4 commits (1 fix), 6 files: “fix(mods): address Greptile findings on \#1038 (\#1039)”; “Remove beta / early access labelling from Mods (\#1083)”; “mods: finish the changelog 2.1.288–2.1.319 plan (catch handlers, ceiling, resume, env holds) (\#1454)”; and 1 more (refs \#1038, \#1039, \#1083, \#1454).
cli-tool/components/mods/integrations/neon-branch-per-session · medium confidence · unverified
4 commits (1 fix) modifying cli-tool/components/mods/ui/agent-flow
A change to existing behaviour in cli-tool/components/mods/ui/agent-flow — 4 commits (1 fix), 6 files: “fix(mods): ignore subagent turns at turn.start (\#1449)”; “Remove beta / early access labelling from Mods (\#1083)”; “mods: finish the changelog 2.1.288–2.1.319 plan (catch handlers, ceiling, resume, env holds) (\#1454)”; and 1 more (refs \#1449, \#1083, \#1454, \#993).
cli-tool/components/mods/ui/agent-flow · medium confidence · unverified
4 commits (1 fix) modifying dashboard/src
A change to existing behaviour in dashboard/src — 4 commits (1 fix), 1 file: “fix(seo): noindex private pages and /api/ instead of blocking them in robots.txt (\#1496)”; “feat(dashboard): real 404s, llms.txt, OpenAPI and trust pages for agents (\#1200)”; “feat(dashboard): CSP in report-only mode, security headers on SSR pages (\#1444)”; and 1 more (refs \#1496, \#1200, \#1444, \#571).
dashboard/src · medium confidence · unverified
5 commits (0 fixes) modifying cloudflare-workers/pulse
A change to existing behaviour in cloudflare-workers/pulse — 5 commits, 5 files: “Remove every Discord integration (\#1446)”; “feat: add Cloudflare Workers for docs-monitor and pulse weekly KPI report”; “Measure returning installers and feature use, and correct /privacy (\#1442)”; and 2 more (refs \#1446, \#1442).
cloudflare-workers/pulse · medium confidence · unverified
5 commits (1 fix) modifying cli-tool/components/skills/scientific
A change to existing behaviour in cli-tool/components/skills/scientific — 5 commits (1 fix), 1030 files: “Update Scanpy settings in SKILL.md (\#411)”; “feat: Add 139 scientific skills from K-Dense-AI/claude-scientific-skills”; “skills: set name to the folder name in 100 skills (\#1269)”; and 2 more (refs \#411, \#1269, \#818, \#677).
cli-tool/components/skills/scientific · medium confidence · unverified
5 commits (2 fixes) modifying cli-tool/components/mods/security/jev-auto-mode
A change to existing behaviour in cli-tool/components/mods/security/jev-auto-mode — 5 commits (2 fixs), 9 files: “feat(mods): add jev-auto-mode, a JSON-policy permission layer judged by Jev (\#995)”; “fix(mods): jev-auto-mode on Claude Code 2.1.307+ (tests and headless engine asks) (\#1367)”; “fix(mods): ignore subagent turns at turn.start (\#1449)”; and 2 more (refs \#995, \#1367, \#1449, \#1083, \#1454).
cli-tool/components/mods/security/jev-auto-mode · medium confidence · unverified
6 commits (0 fixes) modifying cli-tool/components/mods/ui/stack-viewer
A change to existing behaviour in cli-tool/components/mods/ui/stack-viewer — 6 commits, 6 files: “feat(mods): add ui/stack-viewer, a pane mapping the session's components to their scopes (\#1233)”; “feat(mods/stack-viewer): explain an empty built-in skills scope (\#1244)”; “feat(mods/stack-viewer): make scopes read as openable (\#1238)”; and 3 more (refs \#1233, \#1244, \#1238, \#1237, \#1243, …).
(repo-wide) · medium confidence · unverified
6 commits (1 fix) modifying database
A change to existing behaviour in database — 6 commits (1 fix), 5 files: “feat(tracking): record whether an AI agent ran each install (rebased \#1270 + endpoint tests) (\#1368)”; “feat: Add Claude Code changelog monitor with Discord notifications”; “Measure returning installers and feature use, and correct /privacy (\#1442)”; and 3 more (refs \#1270, \#1368, \#1442, \#1170, \#1448).
database · medium confidence · unverified
63 commits (13 fixes) modifying dashboard/src/lib
A change to existing behaviour in dashboard/src/lib — 63 commits (13 fixs), 29 files: “feat(dashboard): hide Loops from navigation menu (\#738)”; “feat: Function Hooks section (experimental) — 10 hooks, CLI flag, blog (\#867)”; “feat: external skills installed from their author's repo (first: t3dotgg/tstack) (\#1279)”; and 60 more (refs \#738, \#867, \#1279, \#850, \#1496, …).
dashboard/src/lib · medium confidence · unverified
7 commits (1 fix) modifying cli-tool/src/analytics-web/services
A change to existing behaviour in cli-tool/src/analytics-web/services — 7 commits (1 fix), 3 files: “feat: Add HeaderComponent for unified header across analytics pages and update version handling”; “feat: migrate analytics dashboard to Astro + modernize TUI”; “feat: Implement Claude Code Console Bridge and Enhanced State Detection (\#14)”; and 4 more (refs \#14, \#1351).
cli-tool/src/analytics-web/services · medium confidence · unverified
87 commits (31 fixes) modifying dashboard/src/components
A change to existing behaviour in dashboard/src/components — 87 commits (31 fixs), 43 files: “feat(dashboard): hide Loops from navigation menu (\#738)”; “feat: Function Hooks section (experimental) — 10 hooks, CLI flag, blog (\#867)”; “fix: Remove marketplace badge from plugin cards”; and 84 more (refs \#738, \#867, \#1518, \#1424, \#1337, …).
dashboard/src/components · medium confidence · unverified
Analytics engine refactored into modular components with new agent and session analysis
The analytics backend has been restructured from a monolithic file into a modular architecture, introducing dedicated classes for conversation loading (ConversationAnalyzer), file watching (FileWatcher), process detection (ProcessDetector), state calculation (StateCalculator), and session analysis (SessionAnalyzer). This refactor includes a new AgentAnalyzer to track specialized agent usage patterns and a YearInReview2025 module for yearly statistics. A multi-level DataCache system has been added to optimize performance and reduce memory usage by caching file content, parsed conversations, and computed results. Additionally, the system now detects both 'Task' and 'Agent' tool names for analytics and provides real-time updates via a NotificationManager. Also in this area: 1 commit (1 fix) fixing cli-tool, cli-tool/analytics-ui.
cli-tool/src/analytics · medium confidence · unverified
Blog content and assets migrated to the dashboard
The blog for www.aitmpl.com has been moved into the dashboard's public directory, establishing a new file-based content system where articles are managed via a central \blog-articles.json\ index and served through a dynamic loader. This change introduces a suite of new SVG cover images for topics including Function Hooks, Agent Teams, and security workflows, and adds the first article page for the Code Reviewer Agent, shifting the blog's hosting location from an external repository to the dashboard's static assets. Also in this area: 11 commits (0 fixes) modifying dashboard; Claude Mods feature is now GA and component content is updated.
dashboard/public/blog · medium confidence · unverified
Dashboard UI overhaul with terminal theme and new content pages
The dashboard has been visually redesigned with a terminal-inspired aesthetic, introducing a dark-mode-first color palette and monospace typography across the interface. This update adds dedicated styling for several new content areas: a blog section with article cards, a component library page with back-navigation and loading states, a featured page with partner badges, and a partnership banner. It also introduces detailed layouts for plugin and stack detail pages, a trending analytics view with growth charts, and a workflow builder with a component tree and modal previews.
dashboard/public/css · high confidence
Dashboard migration to Astro on Cloudflare Pages with new integrations
The dashboard has been rebuilt using Astro with React and Tailwind CSS and migrated from Vercel to Cloudflare Pages. This change introduces support for Clerk authentication, GitHub OAuth for the 'Send to Repo' feature, and Discord bot interactions. It also enables sponsored ad placements (controlled by the PUBLIC\_ADS\_ENABLED flag) and WebMCP catalog tools (controlled by PUBLIC\_WEBMCP\_ENABLED), while configuring the environment for Cloudflare's runtime via wrangler.toml and middleware.
dashboard · high confidence
DoorDash CLI safety and audit hooks
New pre- and post-execution hooks have been added to the DoorDash CLI integration to enforce safety and provide transparency. The allergy-checkout-gate and group-checkout-gate hooks now block checkouts if anaphylaxis-tier allergen matches are detected in the cart or group roster, ensuring deterministic protection against severe allergic reactions. A spend-guard hook prevents budget bypasses by requiring all checkout and cart-addition commands to go through the dd-guard.sh wrapper, which enforces spending policies. Additionally, an audit-log hook passively records every CLI invocation to an append-only JSONL file, allowing users to review command history and costs without altering session behavior.
cli-tool/components/hooks/doordash · high confidence
Enforce Git Flow and Conventional Commits via new hooks
Added three new Git hooks that validate developer actions before execution: conventional commit messages must follow the standard type(scope): description format; direct pushes to protected branches (main, develop) are blocked to enforce pull request workflows; and new branch names must follow Git Flow conventions (feature/, release/v\*, hotfix/). These hooks intercept Bash commands to deny non-compliant operations with helpful error messages and guidance.
cli-tool/components/hooks/git · high confidence
Mods are now generally available and the api-architect agent is updated
Claude Mods are no longer in early access; they are enabled by default in Claude Code 2.1.287 and later, so users no longer need to set the CLAUDE\_CODE\_ENABLE\_FUNCTION\_HOOKS environment variable to use them. The cli-tool also includes an updated api-architect agent that now recommends the graphql-ws protocol over the deprecated subscriptions-transport-ws, suggests Trusted Documents over Automatic Persisted Queries for security, adds ETag/Cache-Control handling for REST clients, updates Apollo Federation guidance to v2.15+, and expands its OWASP API Security Top 10 checklist. Additionally, a new PriceWin MCP server for live hotel and flight price comparison has been added to the web-data components. Also in this area: New UI mods: agent-flow, git-sidebar, and tool-timing-badge.
cli-tool/components/mods/ui/git-sidebar · high confidence
Fixes
1 commit (1 fix) fixing cli-tool, cli-tool/analytics-ui
A fix in cli-tool, cli-tool/analytics-ui — 1 commit (1 fix), 57 files: “CLI 1.31.0: remove unused commands, fix the ones that stay, fix usage tracking (\#1351)” (refs \#1351). Also in this area: New Astro-based analytics dashboard with real-time updates and mobile chat interface.
_cli-tool/analytics-ui, cli-tool/src/analytics-web/\astro · medium confidence · unverified
Test coverage
1 commit adding/updating tests in cli-tool/tests, cli-tool/tests/integration; 23 commits adding/updating tests in dashboard/tests; 8 commits adding/updating tests in cli-tool/tests/unit; New testing infrastructure and CLI tooling for claude-code-templates.
Dependencies
Introduce Rust CLI core and expand sandbox/skill dependency manifests
This change adds a new Rust-based CLI core (\cli-rust\) with its \Cargo.toml\ and \Cargo.lock\, providing a native \cct\ binary alongside the existing Node.js tool. It also introduces dependency manifests for several new sandbox providers (Cloudflare Workers, Docker, and E2B) and various skill scripts (including Playwright, Lighthouse, and Google Analytics), establishing the required package versions for these new execution environments and tooling capabilities.
(dependencies) · high confidence
Written by cca.canine.dev from the codebase's own history, inside the signed delivery this page is composed from.
How this codebase got here
Score
- CAI 42 → 42 (+0.3)
- Rubric changed (rubric-2026.11.2 → rubric-2026.11.4) — scores are not directly comparable.
Lenses
- Code Health 36 → 28 (-8.7)
- Architecture 93 → 94 (+0.1)
- Maturity 85 → 94 (+9.3)
- Readiness 42 → 47 (+4.6)
- Security 40 → 74 (+33.6)
- Accessibility 46 → 47 (+1.3)
- Performance 60 → 60 (+0.0)
Resolved (223)
- Classify ADR enforcement: Architecture Decision Records (cli-tool/components/skills/development/architecture-decision-records/SKILL.md)
- Classify ADR enforcement: Deepspeed - 2020 (cli-tool/components/skills/ai-research/distributed-training-deepspeed/references/2020.md)
- Classify ADR enforcement: Deepspeed - 2023 (cli-tool/components/skills/ai-research/distributed-training-deepspeed/references/2023.md)
- Critical CVE: [GHSA redacted] (cli-tool/package-lock.json)
- Critical CVE: [GHSA redacted] (package-lock.json)
- Critical CVE: [GHSA redacted] (dashboard/package-lock.json)
- Critical CVE: [GHSA redacted] (dashboard/package-lock.json)
- Critical vulnerability: [GHSA redacted] (cli-tool/analytics-ui/package-lock.json)
- Critical vulnerability: [GHSA redacted] (dashboard/package-lock.json)
- Duplicated block (10 lines × 2) (cli-tool/components/skills/business-marketing/app-store-optimization/aso_scorer.py)
- Duplicated block (10 lines × 2) (cli-tool/components/skills/business-marketing/app-store-optimization/metadata_optimizer.py)
- Duplicated block (10 lines × 3) (cli-tool/components/skills/development/skill-development/scripts/aggregate_benchmark.py)
- Duplicated block (10 lines × 3) (cli-tool/components/skills/scientific/pymc/scripts/model_diagnostics.py)
- Duplicated block (11 lines × 3) (cli-tool/components/skills/development/skill-development/scripts/generate_report.py)
- Duplicated block (12 lines × 2) (cli-tool/components/agents/obsidian-ops-team/Scripts/link_suggester.py)
- Duplicated block (12 lines × 2) (cli-tool/components/sandbox/e2b/e2b-monitor.py)
- Duplicated block (12 lines × 2) (cli-tool/components/skills/business-marketing/seo-fundamentals/scripts/seo_checker.py)
- Duplicated block (12 lines × 2) (cli-tool/components/skills/scientific/citation-management/scripts/extract_metadata.py)
- Duplicated block (12 lines × 2) (cli-tool/components/skills/scientific/medchem/scripts/filter_molecules.py)
- Duplicated block (13 lines × 2) (cli-tool/components/skills/development/security-compliance/scripts/risk_calculator.py)
- …and 203 more
New (195)
- CertificationsView.CertificationsView (cognitive 19) (dashboard/src/components/CertificationsView.tsx)
- CertificationsView.CertificationsView (cyclomatic 20) (dashboard/src/components/CertificationsView.tsx)
- CertificationsView.FacetDropdown (cyclomatic 17) (dashboard/src/components/CertificationsView.tsx)
- Critical CVE: [GHSA redacted] (cli-tool/components/skills/ai-research/loki-mode/examples/todo-app-generated/backend/package-lock.json)
- Critical CVE: [GHSA redacted] (cli-tool/components/skills/ai-research/loki-mode/examples/todo-app-generated/backend/package-lock.json)
- Duplicated block (12 lines × 2) (cli-tool/components/skills/scientific/citation-management/scripts/extract_metadata.py)
- Duplicated block (23 lines × 2) (cli-tool/components/skills/scientific/deepchem/scripts/graph_neural_network.py)
- Duplicated block (25 lines × 2) (cli-tool/components/skills/security/security-ownership-map/scripts/build_ownership_map.py)
- Edited copies of one member (3 members, 15+ corresponding lines) (cli-tool/components/skills/scientific/citation-management/scripts/extract_metadata.py)
- Edited copies of one member (3 members, 15+ corresponding lines) (cli-tool/components/skills/scientific/simpy/scripts/resource_monitor.py)
- Edited copies of one member (3 members, 28+ corresponding lines) (cli-tool/components/skills/scientific/gene-database/scripts/fetch_gene_data.py)
- Edited copies of one member (3 members, 42+ corresponding lines) (cli-tool/components/skills/scientific/pytorch-lightning/scripts/quick_trainer_setup.py)
- Edited copies of one member (3 members, 50+ corresponding lines) (cli-tool/components/skills/analytics/google-analytics/scripts/analyze.py)
- Edited copies of one member (3 members, 8+ corresponding lines) (cli-tool/components/hooks/monitoring/context-timeline.py)
- Edited copies of one member (3 members, 8+ corresponding lines) (cli-tool/components/skills/scientific/citation-management/scripts/extract_metadata.py)
- Edited copies of one member (3 members, 9+ corresponding lines) (cli-tool/components/skills/scientific/fda-database/scripts/fda_query.py)
- Edited copies of one member (4 members, 4+ corresponding lines) (cli-tool/components/skills/creative-design/imagegen/scripts/image_gen.py)
- Edited copy of a member (10 corresponding lines) (cli-tool/components/skills/business-marketing/app-store-optimization/aso_scorer.py)
- Edited copy of a member (10 corresponding lines) (cli-tool/components/skills/scientific/gene-database/scripts/batch_gene_lookup.py)
- Edited copy of a member (10 corresponding lines) (cli-tool/components/skills/security/security-ownership-map/scripts/community_maintainers.py)
- …and 175 more
Changes since last survey
- 149 commits — 132 feature/other, 17 fixes
By area
- dashboard/public — 58 commits
- cli-tool/components — 40 commits
- dashboard/src — 27 commits
- (root) — 11 commits
- .github/workflows — 3 commits
- cli-tool/src — 2 commits
- docs/images — 2 commits
- docs/star-history.svg — 2 commits
- .github/ISSUE_TEMPLATE — 1 commit
- cli-tool/README.md — 1 commit
- docs-site/components — 1 commit
- scripts/generate_plugins_json.py — 1 commit
Notable commits
- fix: Fix trending data frozen since RLS: request only granted columns (#1448)
- fix: chore(community): issue/PR templates, SECURITY response times, CODEOWNERS fix, grouped Dependabot (#1400)
- fix: fix(cli): 1.33.1, --skill installs reported failure and were not tracked (#1503)
- fix: fix(cli): only report a real GitHub rate limit as "rate limit" (#1382)
- fix: fix(component-page): stop showing a component's date (#1564)
- fix: fix(dashboard): drop Tailwind antialiased class on body (#1457)
- fix: fix(dashboard): one first collection under concurrent saves; prune stale pending action (#1447)
- fix: fix(dashboard): raise dark-theme text contrast to GitHub/X levels (#1456)
- fix: fix(dashboard): route in-app browser users to Safari/Chrome before Clerk sign-in (#1430)
- fix: fix(dashboard): send CSP reports to Sentry as warnings, one issue per directive + origin (#1468)
- fix: fix(dashboard): stop clipping the sidebar logo tile on hover and when collapsed (#1511)
- fix: fix(funding): serve funding.json from www.aitmpl.com for FLOSS/Fund (#1495)
- fix: fix(mods): ignore subagent turns at turn.start (#1449)
- fix: fix(seo): noindex private pages and /api/ instead of blocking them in robots.txt (#1496)
- fix: fix(seo): one URL per component page, crawlable links, honest sitemap (#1388)
- fix: fix(skills): drop render-internal.com from render-deploy (#1466)
- fix: security: npm audit fix, CodeQL scoped to our code, retire catalog-wide SkillSpector alerts (#1406)
- change: 1.32.2
- change: Add AsoBeast MCP to marketing (#1387)
- change: Add BrewPage MCP to devtools (#1439)
- …and 129 more
Architecture
- Unchanged — 0 containers · 2 contexts · 0 edges
Written by cca.canine.dev from the codebase's own history, inside the signed delivery this page is composed from.
Survey your own repository
davila7/claude-code-templates was measured the same way every project in this corpus was: the same rubric, at a pinned commit, with the result published in full. Point a surveyor at a repository you know and see whether you agree with it.
About this page
- The score is its most recent published measurement, taken on 16 November 2026 at a pinned commit. It is not a live figure and does not change until the project is measured again.
- Measured at commit 1530598131e773d2dc1bdefa7d28df64a7d6b656 — the exact code this score is about.
- Scored under rubric-2026.11.4 — the same rubric and the same method as every other entry in this index.
- Measured by Canine Code Analyzer (cca.canine.dev) using cca measure-2026.11.73+preprod-52de90b74216.