seaworld008/commonly-used-high-value-skills

A high-value Skills repository for global AI developers and Agent workflow builders, covering high-frequency task scenarios such as development engineering, DevOps, product design, operations, office automation, financial investment, AI platforms, and security governance. Currently, there are 16 categories / 294 skills.

v4.1.0 Upstream Contract and Cross-Platform Fix

Updated engineering methods, Feishu tasks/charts, AKS readiness, and Neon search guidance, and fixed macOS offline HTML report updates. Live Panel retains rendering capabilities and original license text, replacing third-party diagram configurations with original simulation examples. View release notes.

v4.0.0 Security Reinforcement

Addressed the CodeQL risks introduced in HTML/Live Panel implementation in v3.3.0, and added access control for the publishing link to “Main branch open Code Scanning alert must reset to zero.” HTML CLI switched from high-risk upstream bundled scripts to restricted offline implementation, with some old features explicitly disabled. 294 Skills and 16 categories retained.

v3.3.0 Featured Enhancements

Go trio, architecture refactoring, HTML visualization, dynamic architecture panel, and Azure diagnostics; Upgrade existing Agent Browser, fully fix upstream sources and comply with authorization boundaries.

Why It’s Worth Collecting

• Collect all high-frequency reusable skills at once, reducing the time spent searching for prompts, scripts, and workflows.
• Compatible with multiple AI tool usage modes such as Codex, Claude Code, Hermes Agent, and [T3]].
• Organized by scenario classification, suitable for daily searches, secondary expansion, and team accumulation.
• Many skills are not just documents; they also include scripts/, references/, and assets/, which can be directly reused.
• The warehouse already has automated chains for discovering new skills, synchronizing upstream updates, candidate preferences, quality verification, and view generation, making it suitable for continuous operations rather than one-time collection.
• Now supports unified governance of whitelist sources, blacklist sources, priority sources, and basic thresholds using policy files, making preferred results more stable and controllable.
• scripts/sync_codex_skills.py is now available, allowing you to sync the latest skills in your warehouse to [T8]], Claude Code and other local skill directories with one click, reducing manual copying and version drift.
• In addition to functional coverage, warehouses also emphasize security and trustworthiness: they offer source tracking, candidate screening, and pre-installation risk identification capabilities, as well as built-in security review skills such as skill-vetter, skill-security-auditor, input-guard, and [T13].
• License auditing and monthly deadlink inspections are now built-in: repo-validation blocks external skills missing license metadata from entering the main branch, and dead-links workflows generate monthly backlink inspection reports.