vrt-classifier

Classify security findings against the Bugcrowd VRT taxonomy with priority, CWE, and CVSS baselines.

7|4|Updated Jun 22, 2026
One-click install
npx skills add https://github.com/dbx0/skills --skill vrt-classifier-dbx0
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: vrt-classifier
Source: https://github.com/dbx0/skills/tree/main/skills/methodology/triage/vrt-classifier
Command: npx skills add https://github.com/dbx0/skills --skill vrt-classifier-dbx0

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve? Bug bounty reports need a defensible severity rating, and picking the wrong Bugcrowd VRT entry or inflating a priority gets submissions bounced or downgraded. This Skill maps a demonstrated finding to exactly one VRT leaf entry and carries its baseline priority (P1-P5), CWE, and CVSS v3 vector into the report. ## Core Features & Use Cases - Taxonomy Search and Lookup: Query a local mirror of the official bugcrowd/vulnerability-rating-taxonomy release (437 leaf entries across 26 categories) with scripts/vrt.py using search, show, list, and flat commands. - Priority Adjustment Guidance: Apply documented upgrade/downgrade modifiers (tenant boundary crossings, preconditions, self-only impact) off the baseline with stated reasons. - Structured Output Contract: Produce a report-ready tuple of VRT id-path, priority, CWE, baseline and as-demonstrated CVSS vectors, plus Bugcrowd's own remediation advice. - Use Case: You demonstrated an IDOR exposing another tenant's invoices. Search the taxonomy, descend to the exact leaf variant, record the P1 baseline with CWE-932 and the CVSS vector, and paste the remediation text into your submission. ## Quick Start Classify my finding of an unauthenticated IDOR that exposed other users' invoice PDFs against the Bugcrowd VRT and give me the priority, CWE, and CVSS vector for the report.

Frequently Asked Questions about vrt-classifier

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

FAQPage Schema
How do I classify a bug bounty finding with the Bugcrowd VRT?

State the demonstrated impact in one sentence, then run scripts/vrt.py search with keywords for the mechanism and impact. Descend to a leaf entry with show, apply priority modifiers, and record the id-path, priority, CWE, and CVSS vector.

What priority should I assign a vulnerability for Bugcrowd?

Use the baseline priority of the matching VRT leaf entry, then adjust only for documented modifiers such as crossing a tenant boundary or requiring unrealistic preconditions. Every deviation from baseline must be stated with its reason in the report.

Does the VRT classifier work for programs with custom severity tables?

No. A program brief's own severity table or out-of-scope list overrides the VRT, so read the brief first. The classifier is for Bugcrowd or VRT-aligned programs and falls back to plain CVSS when no VRT analogue exists.

Why does my VRT search return no results for common bug names?

VRT names are formal and diverge from hunter shorthand; for example, searching s3 bucket fails while misconfigured cloud storage matches. Retry with taxonomy vocabulary or list the plausible category and its children before concluding no entry exists.

How current is the VRT data used for classification?

The references directory mirrors the official bugcrowd/vulnerability-rating-taxonomy repository pinned at release 2026-07-08. When a classification hinges on a recently added category, re-pull the upstream repo to avoid stale data.