AI-Tooling Guide
Cross-team Claude workflow programme at JustGiving: patterns and plugins shared across teams.
Why
During 2025 my work moved from using AI tools to designing patterns other engineers could adopt. Every team I worked alongside was rediscovering the same prompt patterns, plugin shapes, and agent-mode trade-offs on its own.
A shared evaluation hub gives teams one place to record which patterns held up and which did not, so the next engineer starts from a tested chatmode or skill instead of a blank prompt.
How it’s built
Contributed to the internal cross-team `ai-coding-tools-explorer` repository alongside AI Champions: spec-driven workflow patterns, discovery-focused chatmodes, data-diagram-focused agents, runbook generation, plugin examples, and lessons from my own tooling, written up so other engineers can reuse them.
Outside work, my personal examples include a Claude MCP server marketplace and an OpenCode-backed stdio MCP server for asynchronous agent tasks. I also run bi-weekly AI tooling sessions and have the DataCamp Associate AI Engineer track in progress.
- Shared evaluation hub for AI coding tools: patterns, plugins, learnings
- Spec-driven workflow templates other engineers can fork
- Personal MCP examples: a Claude server marketplace and an OpenCode-backed asynchronous agent server
- Bi-weekly AI tooling meetings for cross-team sharing
- DataCamp Associate AI Engineer track alongside the practical tooling work
- Discovery-focused chatmodes and data-diagram-focused agents documented as patterns
Demo
Screenshots
Links
Active. Source for the employer artefacts is closed; the patterns are available to walk through in conversation. Screenshots below show the AI tools explorer with representative, non-confidential content.