What the harness adds
around a scanner.
A scanner reports what it saw. BountyHarness adds scope, planning, safety gates, impact verification, and the evidence needed for another person to reproduce the result.
| BountyHarness | Single-purpose scanner | Ad-hoc scripts | |
|---|---|---|---|
| Bug classes | 46 skills | 1 (e.g. nuclei) | Whatever you wrote last |
| Evidence standard | Request + response + PoC + impact | Raw output | Inconsistent |
| Safety gating | 4 tiers + scope + circuit breaker | None | Manual discipline |
| Tool management | 99 tools, capability-mapped | Bundled templates | Manual installs |
| Planning | Domain-driven ranked plan | Run everything | Gut feel |
| Knowledge reuse | Program memory, asset graph | None | Notes files |
| Autonomous mode | bb-hunt from one URL | Manual chaining | Cron + hope |
| Report quality | Impact-verified, CVSS, platform export | Not a report | Copy-paste |
When to use what
Use BountyHarness when
You want a repeatable, evidence backed workflow across many bug classes, especially when an agent is driving.
Use a scanner when
You need fast, broad coverage for a single class (e.g. nuclei for CVE templates). BountyHarness actually wraps nuclei via the nuclei-scanner skill.
Not either/or. BountyHarness orchestrates scanners. It does not replace them. The
recon, nuclei-scanner, and cloud skills wrap and scope the same tools you'd run manually, with rate limits and evidence capture built in.