- Before adressing any task, ensure you have read
agents/guides/workflow.mdand the relevant templates inagents/guides/workflow-extras/. - Always review the task's
spec.mdandimplementation.md(if present) to understand the scope and requirements. If these documents are missing or unclear, seek clarification before proceeding. - Implementation logs should include a step-by-step checklist capturing verification of 100% pytest coverage and resolution of all Ruff findings prior to completion.
agents/guides/: Core reference material on engineering practice, patterns, style, and workflow.agents/tasks/feat/: Feature work specs, each folder numbered with a slug and containingspec.mdplus follow-on artifacts.agents/tasks/patch/: Patch or follow-up efforts (refactors, hardening), organized the same way.
agents/guides/engineering-guide.md: Micro-design principles covering seam-first architecture, dependency injection, testing focus, and other day-to-day engineering defaults.agents/guides/patterns-and-architecture.md: Deep dive on organization patterns, composition vs. inheritance guidance, module layout, logging, and anti-patterns.agents/guides/styleguides.md: Language and tooling conventions (Python, docs, front-end snippets), docstring guidance, Ruff configuration, pytest patterns, and semantic line break practices.agents/guides/workflow.md: Canonical collaboration loop with the human, including task lifecycle, templates, and expectations for history updates and reviews.agents/guides/workflow-extras/: Template library used byworkflow.md.codereview-tpl.md: Structure for optional review deliverables.implementation(-mini)-tpl.md: Full and lightweight execution logs.spec(-mini)-tpl.md: Full and lightweight task specs.
Use these references when scoping new work, reviewing deliverables, or aligning implementation details with existing expectations.