KKKKhazix/khazix-skills

This system is a multi-workspace repository that aggregates a diverse set of independent microservices and agent skills, ranging from infrastructure utilities like a link shortener and PDF splitter to specialized AI capabilities for news aggregation and deep research. It also includes operational tools for project governance, storage analysis, and authentication, alongside foundational API contracts and task management services. The codebase serves as a hub for both backend service development and AI agent skill definitions, supported by dependency manifests for various evaluation projects.

Users can now install the AIHOT Agent Skill to query current AI news, curated highlights, hot topics, and daily digests via the AIHOT public API. The skill provides a structured workflow for agents to fetch data anonymously and read-only from the stable v1 API endpoints (e.g., \/api/v1/items\, \/api/v1/hot-topics\, \/api/v1/dailies\) without requiring an API key. It includes a bash installer (\install.sh\) that supports multiple agent targets (Claude, Codex, Gemini, Copilot, OpenCode) and handles local state management via a random \aihot-actor-id\ for cross-channel deduplication. The skill enforces specific routing rules, such as defaulting to the 'selected' mode for general queries and falling back to 'all' for keyword searches, while ensuring output is formatted as concise Chinese briefs with proper attribution and time-window handling.

aihot · high confidence

Adds the neat-freak skill (v3.0.0), an agent capability that reconciles project documentation, rules (CLAUDE.md/AGENTS.md), authorized memory, and workspace residue with the actual code and runtime state. It provides a lightweight path for small projects and a full path for complex releases, enforcing strict scope boundaries, safe memory handling (including read-only generated memory), and a two-stage reporting process that requires explicit user confirmation before destructive cleanup. The skill is supported by a comprehensive evaluation suite (11 behavioral tests) and a read-only inventory script to verify structural integrity and trigger contracts.

neat-freak · high confidence

This change introduces several new application components and services across different workspaces. In the taskflow workspace, a tRPC-based API server is added with endpoints for managing tasks (list, create, update, delete) using a standalone HTTP and WebSocket server on port 4000. The notesapp workspace now includes authentication middleware using Clerk to protect specific routes. An analytics dashboard is introduced in the analytics\_dashboard workspace, featuring a React component that displays page event data via a line chart, backed by a Supabase client. The auth-center workspace adds OAuth 2.0 Device Authorization Grant (RFC 8628) routes for device code issuance, token polling, and session management, along with a corresponding DevicesPage component in the skills-hub workspace to display authorized devices. Additionally, a simple link shortener service is added in the Link\_Shortener workspace using Express, and a PDF splitting utility is introduced in the pdf-tools workspace using FastAPI and pypdf. Finally, a GraphQL API contract is defined in the current-app workspace.

(repo-wide) · high confidence

A new storage-analyzer skill has been added that performs read-only disk usage scans on both macOS and Windows, classifies large directories into three safety tiers (auto-clean, manual review, and caution), and generates an interactive HTML report. The report includes a visual breakdown of disk usage and provides one-click actions—such as moving files to the Trash or deleting them directly—via a local web server, while strictly enforcing a whitelist-based security model that only allows operations on paths explicitly identified as safe for deletion.

storage-analyzer · high confidence

A new 'hv-analysis' skill has been added to enable deep, systematic research on products, companies, concepts, or individuals using a dual-axis framework. This capability guides the user through longitudinal historical analysis and cross-sectional competitive benchmarking to produce a comprehensive, narrative-driven PDF report. The implementation includes a detailed skill definition (SKILL.md) outlining the research methodology, a structured data schema (references/schema.json) for organizing the analysis, and a Python script (scripts/md\_to\_pdf.py) that converts the final Markdown report into a professionally formatted PDF using WeasyPrint.

hv-analysis · high confidence

New package.json and requirements.txt files have been added to several evaluation workspace directories (taskflow, quicktodo, analytics-dashboard, link-shortener, current-app, and the unknown-platform project). These manifests define the runtime and development dependencies for the respective projects, including libraries such as React 19, Express, Flask, tRPC, Zod, and GraphQL Yoga.

(dependencies) · high confidence

Written by watchdog.canine.dev from the codebase's own history, inside the signed delivery this page is composed from.

• First survey — no prior run to compare against. CAI 49.

• Code Health 98
• Architecture 69
• Maturity 60
• Readiness 24
• Security 100

• 76 commits — 71 feature/other, 5 fixes

• (root) — 33 commits
• aihot/README.md — 12 commits
• aihot/SKILL.md — 11 commits
• aihot/references — 7 commits
• neat-freak/SKILL.md — 3 commits
• kaizike-writer/SKILL.md — 2 commits
• neat-freak/references — 2 commits
• vibeguard/scripts — 2 commits
• aihot/.gitignore — 1 commit
• khazix-writer/references — 1 commit
• neat-freak/evals — 1 commit
• storage-analyzer/scripts — 1 commit

• fix: Fix Codex skills path in neat-freak references and update CLAUDE.md
• fix: Fix skill installation path to ~/.claude/skills/
• fix: Revert SKILL.md to original state
• fix: fix(aihot): repair the install command, sync the capability list
• fix: fix(readme): unbreak bold markers and fit the seven-questions table
• change: Add English README and language switcher
• change: Add WeChat article links to README tables
• change: Add author field and credit line to kaizike-writer skill
• change: Add hv-analysis skill for horizontal-vertical deep research
• change: Add neat-freak skill for end-of-session knowledge cleanup
• change: Add skill installation paths for all supported tools
• change: Add 横纵分析法 deep research prompt and update README
• change: Initial release: open-source kaizike-writer skill
• change: Link neat-freak article and simplify install to agent-only method
• change: Redesign README as hero showcase with per-skill cards
• change: Reframe README as AI toolkit covering both prompts and skills
• change: Remove OpenCode from supported tools list
• change: Replace kaizike-writer with updated khazix-writer skill
• change: Restore 讲解 column with WeChat article links in skill list tables
• change: Rewrite README in plainer voice and surface skill list above install
• …and 56 more

Written by watchdog.canine.dev from the codebase's own history, inside the signed delivery this page is composed from.

KKKKhazix/khazix-skills 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.