bilinote

Read the backend development guidelines before starting your development task.

Read the frontend development guidelines before starting your development task.

Deep requirements and design discovery workflow that keeps Trellis task-first PRD capture but adds stronger design confirmation, explicit technical design notes, and a clean handoff to task plan generation. Use when you want a richer brainstorming experience than $brainstorm, when trade-offs need deliberate review, or when you want Superpowers-style design validation inside Trellis tasks.

Check if the code you just wrote follows the backend development guidelines.

Cross-Layer Check

Check if the code you just wrote follows the frontend development guidelines.

指导 agent 何时以及如何更新任务的 .fusion/ 执行态文件

Create New Skill

Execute the current task's plan.md slice by slice using strict red-green-refactor discipline. Reads the task PRD, technical design, and execution plan, reviews for gaps, and then implements only through failing tests first. Use after $write-task-plan and before finish-work.

Finish Work - Pre-Commit Checklist

Synthesize durable lessons from the current task and promote them into .trellis/spec/ without polluting specs with one-off task noise. Reads the task artifacts, changed files, tests, and design decisions, then updates the right spec documents or explicitly decides that no durable spec update is needed. Use after implementation and before finish-work.

Integrate Skill into Project Guidelines

PART 3: Customize Your Development Guidelines

Record work progress after human has tested and committed code

Two-stage cross-review using subagents: spec compliance review first (does code match PRD?), then code quality review (is it well-built?). Each stage uses an independent subagent with isolated context to catch blind spots the implementer missed. Use after all execute-plan-tdd slices are complete for complex or high-risk tasks where single-agent self-review is not enough.

Four-phase debugging methodology: root cause investigation, pattern analysis, hypothesis testing, and TDD-based fix implementation. Enforces evidence-first discipline and stops random fix attempts. Use when encountering any bug, test failure, or unexpected behavior — especially when under time pressure or after a fix attempt has already failed.

Generate a task-local execution plan in .trellis/tasks/<task>/plan.md from the current PRD, technical design, and project specs. Produces small, TDD-first implementation slices with exact files, checks, and sequencing. Use after brainstorming is approved and before implementation begins.

Create, analyze, proofread, and modify Office documents (.docx, .xlsx, .pptx) using the officecli CLI tool. Use when the user wants to create, inspect, check formatting, find issues, add charts, or modify Office documents.

Save current task execution state to .fusion/ for context recovery across sessions. Use at key milestones, before context compaction, or when switching tasks.

Semantically restore current task execution state from .fusion/ recovery data. Use at the start of a new session or after context compaction.