Skip to content
Development
Skill

/gstack-investigate

Systematic root-cause debugging for OPC work. Use when there is a bug, crash, regression, failing command, flaky behavior, provider issue, or unexplained output. Requires evidence before fixes.

From plugin
open-claudecode
93415 skills14 agents14 commands1 MCP
Install
$ npx -y skills add LING71671/Open-ClaudeCode --skill gstack-investigate --agent claude-code

How it fires

How this skill gets triggered: by you, by Claude, or both.

  • Fires itselfAuto-invocation. Claude auto-loads it when your prompt matches the work.Auto-invocation is when the right skill fires by itself at the right moment, driven by a FLOW.md router and a hook, instead of you invoking it by name. It is the difference between a skill being installed and a skill actually getting used.Read the full definition →
  • You can call itInvoke it directly when you want it.
  • Slash command/gstack-investigate

Context preview

The summary Claude sees to decide when to auto-load this skill.

Systematic root-cause debugging for OPC work. Use when there is a bug, crash, regression, failing command, flaky behavior, provider issue, or unexplained output. Requires evidence before fixes.

SKILL.md

gstack-investigate.SKILL.md
name: gstack-investigate
description: Systematic root-cause debugging for OPC work. Use when there is a bug, crash, regression, failing command, flaky behavior, provider issue, or unexplained output. Requires evidence before fixes.
allowed-tools:
  - Read
  - Grep
  - Glob
  - Bash
  - Edit
  - Write

gstack-investigate

Use this skill to diagnose before fixing.

Workflow

1. State the observed symptom and the expected behavior. 2. Reproduce the issue with the smallest command, UI flow, fixture, or log slice available. 3. Trace the relevant path through the code before proposing a fix. 4. Form at least three plausible causes. 5. Test the causes with targeted checks, logs, assertions, or isolated commands. 6. Pick the cause with the strongest evidence. 7. Apply the smallest fix that addresses that cause. 8. Re-run the original reproduction and one nearby regression check.

Rules

  • Do not fix before reproduction or strong evidence.
  • If one attempted fix fails, record why and try a different strategy.
  • Prefer direct runtime evidence: command output, logs, API responses, persisted state, or failing tests.
  • Keep unrelated refactors out of the fix.

Output

Report:

  • Symptom
  • Root cause
  • Fix
  • Verification command or manual check
  • Residual risk
Ships withopen-claudecode

完整开源的 Claude Code 项目 - 基于 Anthropic 官方源码重建 🌐 Languages: 中文 | English

Get the whole plugin