Skip to content
Back to skills

Triage Jordangunn Skills

ASecurity

Perform breadth-first hypothesis surfacing and prioritization across all ownership zones, without commitment. This is not diagnosis and not investigation.

  • 666 stars
  • 0 votes
  • 0 copies
  • 0 views
  • Added September 2, 2026
devopsgokubernetesfrontendbackendci/cd

Security analysis

A100/100

Scanned September 2, 2026

npx -y skills add majiayu000/claude-skill-registry --skill triage-jordangunn-skills --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Triage Jordangunn Skills?

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

Security grade badge for Triage Jordangunn Skills
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/majiayu000-triage-jordangunn-skills/badge)](https://www.skillsdirectory.com/skills/majiayu000-triage-jordangunn-skills)

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: doctor-triage
license: MIT
description: |
  Perform breadth-first hypothesis surfacing and prioritization across all
  ownership zones, without commitment. This is not diagnosis and not investigation.
metadata:
  author: Jordan Godau
  references:
    - 01_PURPOSE.md
    - 02_SCOPE.md
    - 03_BEHAVIOR.md
    - 04_OUTPUT.md
  assets:
    - TRIAGE_REPORT.md
  scripts:
    - evidence.sh
  keywords:
    - triage
    - hypothesis
    - breadth
    - prioritization
    - zones
    - likelihood
---

# doctor-triage

Perform **breadth-first hypothesis surfacing and prioritization** across all ownership zones.

## Purpose

Assume the problem may live in a layer the user is ignoring. Prefer "possible and boring" over "clever and narrow."

This is **not** diagnosis and **not** investigation.

## Epistemic Stance

- The problem may originate outside the layer where symptoms manifest
- Early confidence is a warning sign, not a virtue
- Multiple hypotheses are better than one "obvious" answer

## Scope of Consideration (Always Include)

- Backend application code
- Frontend build/runtime
- CI/CD pipelines
- Container build & images
- Kubernetes manifests (ingress, services, probes, RBAC, secrets)
- Cloud infrastructure (IaC, identity, networking, storage)
- Configuration & environment variables
- Dependencies and versions

## You MUST

- Enumerate plausible causes across ALL zones
- Rank them by likelihood with explanation
- Cite lightweight evidence pointers (file paths, grep hits, manifest locations)
- Explicitly call out assumptions under dispute
- Recommend which suspect(s) merit focused exam next

## You MUST NOT

- Investigate deeply
- Open large files unnecessarily
- Collapse uncertainty into a single answer
- Propose fixes

## Output

A completed **Triage Report** using the template at `../.resources/assets/TRIAGE_REPORT.md`.

## Scripts

- `evidence.sh` — Lightweight grep-based evidence pointer gathering

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.