A small collection of Claude Code skills focused on production AI workflow optimization, with one active skill for hierarchical model usage and cost/time savings. Source is a GitHub repo created and updated in July 2026, no license listed at top level.
Collecting history — the radar snapshots this repo daily. The trend line appears after 3 days of data (1 so far).
What it is
Claude Code Skills is a collection of production-ready Claude Code skills for AI workflow optimization. The README highlights a Multi-Tier Models skill designed for cost-optimized hierarchical model usage.
How it works
The Multi-Tier Models approach combines multiple components to optimize reasoning, planning, and execution:
- Fable 5 (strategic advisor) for high-level reasoning
- Opus 4.8 (orchestrator) for planning and synthesis
- Haiku 4.5 (workers) for text/code execution
- Gemini 3.5 Flash (workers) for multimodal analysis The stated results include cost savings, time reductions, and token reductions, but these are described as real results rather than guarantees for all use cases.
Getting started
Two installation options are provided:
- Install to Claude Code
# Clone this repository
git clone https://github.com/upasana1105/claude-code-skills.git
# Install a skill (copy to your skills directory)
cp -r claude-code-skills/multi-tier-models ~/.claude/skills/
# Use in Claude Code
/multi-tier-models
- Manual Setup
- Navigate to
~/.claude/skills/ - Create a directory for the skill (e.g.,
multi-tier-models) - Copy the skill files from this repository
- Invoke via
/skill-namein Claude Code
Recent releases
- none
Traction
Stars: 14 (no daily/weekly breakdown provided). Forks: 5. Open issues: 1.
License / Caveats
License: MIT License - See individual skill directories for details. The top-level README lists MIT, but the repository metadata here notes "none listed" for license. There is 1 open issue. Created and last pushed on 2026-07-15. No specific age beyond creation date provided in the facts.