Enforce spec-first ADK agent development with adk eval workflows.
npx skills add https://github.com/SeanChenR/maestro-agent --skill adk-dev-guide
Please help me install this Agent Skill. Skill: adk-dev-guide Source: https://github.com/SeanChenR/maestro-agent/tree/main/.agents/skills/adk-dev-guide Command: npx skills add https://github.com/SeanChenR/maestro-agent --skill adk-dev-guide
Provides a single, prescriptive development workflow and mandatory coding guidelines that prevent unintended changes, reduce iteration waste, and ensure ADK agents meet spec-driven quality and evaluation thresholds.
Read the project's DESIGN_SPEC.md, then implement changes with surgical precision, run adk eval for iterative evaluation, and obtain explicit human approval before deploying.
High-intent search queries and answers about installing and using this skill.
To evaluate an ADK agent during development, run iterative adk eval workflows starting with small test cases. Expand test coverage only after core cases pass, ensuring the agent meets spec-driven quality thresholds.
Spec-driven ADK agent development uses DESIGN_SPEC.md as the primary source of truth for implementation decisions. It enforces strict code preservation rules and surgical edits to prevent unintended changes to configuration and model settings.
Troubleshoot ADK model location issues by applying model selection policies and verifying tool import precision. The development workflow includes guardrails for location troubleshooting and loop-breaking heuristics to resolve these failures.
Yes, explicit human approval is required before deploying an ADK agent. The development workflow mandates this deployment guardrail alongside model selection rules and stepwise evaluation procedures to ensure safe deploy pipelines.
The best way to preserve code when editing ADK agent logic is applying surgical edit practices so only targeted lines change. This prevents unintended modifications to unrelated code, configuration, and model settings during implementation.
ADK agent evaluation may fail on edge cases if iterative workflows expand test coverage too quickly. The mandated stepwise evaluation procedure requires starting small and only expanding after core cases pass to isolate failures.
hq-opensourceEnforce a Phase 1-4 ADK agent development lifecycle with DESIGN_SPEC.md.
MinhHoangDonoBuild AI agent applications with Google ADK Python using tools, multi-agent composition, and workflows.
AbdullahMalik17Builds AI agents with the OpenAI Agents SDK using async patterns, tools, and multi-agent handoffs.
JFrangelDesign scalable agent systems with Agent Harness and Context Engineering.
