← All creators

Abolfazlshahi

User

A procedural Agent Skill that teaches custom-provider models how to discover, use, verify, and recover from Claude Desktop/Cowork-style tools and workflows

1 indexed · 0 Featured · 15 stars · avg score 86
View on GitHub →

Categories

AI & Automation (1)

Indexed Skills (1)

AI & Automation Solid

claude-capability-bridge

Teaches third-party and custom-provider models how to operate inside Claude Desktop-like Agent Skills runtimes by discovering real capabilities, selecting and sequencing tools, using browser/Chrome/computer/MCP/connectors/files/Git/code, handling Projects/Skills/Plugins/artifacts/subagents/scheduled work, tracking session and local-vs-cloud state, diagnosing gateway/custom-provider feature gaps, recognizing project execution models before launch, verifying outcomes, and recovering safely from failures. Use for coding, web-app testing, GUI automation, research, file work, integrations, and complex multi-step desktop tasks.

15 Updated yesterday
Abolfazlshahi

Bio shown is the top-scored skill's repo description as a fallback — real GitHub bios land in a future update.