Skip to content
Back to skills

Debug

ASecurity

Diagnose the current omd session or repo state using mode state, notepad, logs, and focused reproduction

  • 8 stars
  • 0 votes
  • 0 copies
  • 2 views
  • Added October 13, 2026
ai-agentsgogit

Works with

  • mcp

Security analysis

A100/100

Pro scans all 2 files and shows the line behind each finding

Scanned October 13, 2026

npx -y skills add 1225Sakura/oh-my-deepseekharness --skill debug --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Debug?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Debug
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/1225sakura-debug/badge)](https://www.skillsdirectory.com/skills/1225sakura-debug)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: debug
description: Diagnose the current omd session or repo state using mode state, notepad, logs, and focused reproduction
when-to-use: The user wants help diagnosing a current omd/DeepSeek-Harness session problem, workflow breakage, or confusing runtime behavior — stuck modes, stale state, orchestration misbehavior. Not for prescribing fixes before the failure is isolated.
---

# Debug

Use this skill when the user wants help diagnosing a current omd/DeepSeek-Harness session problem, workflow breakage, or confusing runtime behavior.

## Goal

Find the real failure signal quickly and explain the next corrective step.

## Workflow

1. Read the user's issue description carefully.
2. Inspect the most relevant local evidence first:
   - mode state: `mcp__omd-state__state_get_status({ cwd })` and `state_list_active({ cwd })`; read one mode with `state_read({ cwd, sessionId, mode })`
   - when MCP is degraded, read `.omd/` artifacts directly with `glob`/`read` — state files live at `.omd/state/sessions/{sessionId}/<mode>-state.json`
   - `mcp__omd-state__notepad_read({ cwd })` when session memory matters
   - failing tests or commands — rerun the narrowest one with `pwsh`
   - `git log` / `git_diff` for recent-change correlation
3. Reproduce the issue narrowly if possible.
4. Distinguish symptoms from root cause.
5. Recommend the smallest next fix or verification step.

## Rules

- Prefer real evidence over guesses.
- Use the state surfaces when the issue involves orchestration, modes, or agent flow.
- OMC's original trace MCP tools (`trace_timeline` / `trace_summary`) are **omd phase-2**; until then, reconstruct agent flow from `.omd/` state and handoff files, `git log`, and focused `grep` over logs.
- If the issue is actually a product/runtime bug (dsh host itself, not omd or app code), say so plainly.
- Do not prescribe broad rewrites before isolating the failure.

## Output

- Observed failure
- Root-cause hypothesis
- Evidence for that hypothesis
- Smallest next action

## State Contract (状态契约)

Debug **holds no mode state**: it is a read-mostly diagnostic lane and creates nothing under `.omd/state/`. Findings worth keeping go to `mcp__omd-state__notepad_write_working`; if a fix follows, the enclosing mode (if any) owns persistence. For deeper causal hunts with competing hypotheses, escalate to the `trace` skill.

Files in this skill

  • SKILL.md2.3 KB
  • SKILL.zh.md2.1 KB

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.