sb-debug

Diagnose why a SpecBind task failed and return a categorized root cause.

Updated Jul 17, 2026
One-click install
npx skills add https://github.com/Huruikagi/pc-build-planner --skill sb-debug-huruikagi
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: sb-debug
Source: https://github.com/Huruikagi/pc-build-planner/tree/main/.agents/skills/sb-debug
Command: npx skills add https://github.com/Huruikagi/pc-build-planner --skill sb-debug-huruikagi

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? When a SpecBind workflow task fails or cannot be implemented, teams often guess at the cause or retry with the same flawed reasoning. This Skill performs a read-only investigation that establishes the true root cause, categorizes it, and routes it to the right owner instead of letting failed attempts repeat. ## Core Features & Use Cases - Structured diagnosis contract: Every run ends with a parseable block naming CATEGORY (IMPLEMENTATION, REVIEW, PLAN, ARTIFACT, ENVIRONMENT, or UNDETERMINED), CAUSE, NEXT_ACTION, and UNCERTAIN items. - Evidence-based investigation: Reads specs, tasks, contracts, designs, implementation notes, and steering documents via the specbind CLI to find where actual behavior first diverged from approved artifacts. - Read-only guarantee: Never modifies tracked files or applies the fix, preserving evidence for whoever implements the correction. - Use Case: A task in your milestone is blocked after two failed attempts. Dispatch this Skill in a fresh context with the failure details; it identifies that the design artifact contradicts the requirements (ARTIFACT category) and routes the fix back to planning instead of another doomed implementation attempt. ## Quick Start Ask the assistant to use sb-debug to diagnose why the current SpecBind task failed and return the categorized diagnosis.

Frequently Asked Questions about sb-debug

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I diagnose why a SpecBind task failed?▼

Dispatch sb-debug in a fresh context with the failure description, spec, and task identity. It reads the approved artifacts via the specbind CLI, finds where behavior first diverged, and returns a categorized diagnosis with a next action.

What failure categories does SpecBind debugging return?▼

Diagnoses are categorized as IMPLEMENTATION, REVIEW, PLAN, ARTIFACT, ENVIRONMENT, or UNDETERMINED. The category determines who owns the fix, so misrouting a design contradiction as an implementation bug is explicitly avoided.

Can sb-debug apply the fix it finds?▼

No. The Skill is strictly read-only: it may run commands that reproduce or observe the failure but never modifies tracked files or applies fixes. The diagnosis is handed to whoever owns the identified category.

Why run failure diagnosis in a fresh context?▼

A retry that inherits the reasoning which just failed reliably reproduces the same failure. A clean context forces re-derivation from evidence rather than from conclusions the failed session already committed to.

What happens when the failure cause cannot be determined?▼

The Skill returns UNDETERMINED and names the specific evidence-gathering step that would distinguish the remaining possibilities. An honest fork between two causes is preferred over a confident but wrong single answer.