BOUNTYHARNESS / RESEARCH TOOLKITOPEN SOURCE · v3.1

For the work between a hunch and a finding.

Follow the lead.
Keep the proof.

Run security research with explicit scope, repeatable workflows, and evidence you can hand to another human.

Explore a run
Conceptual evidence graph
01 / DEFINEScopescope.txt
02 / INVESTIGATEExecuteskill.yaml
03 / SUBSTANTIATEVerifyimpact-verifier
04 / HAND OFFEvidencerequest.txt + poc.sh
46 skill packages99 registered tools4 safety tiersMIT / macOS + Linux

01 / INSIDE A RUN

The steps are visible.
So is what they leave behind.

A walkthrough of the documented workflow.
Illustrative output. No live target requests.

RESEARCH / example.comLOCAL WORKFLOW
~/research/example
$ bb-init example.com --program example --scope-file scope.txt
$ bb-validate

01 Read the target and scope file.

02 Set the output directory and rate limit.

03 Validate the context before execution.

STAGE 01 / 04

02 / THE TOOLKIT

Small core.
Deep skill library.

The harness handles context, dispatch, and traces. Each skill supplies its own workflows, scripts, and research method.

Browse all 46 skill packages

03 / EXPLICIT BOUNDARIES

Give the run
a scope.

Use BountyHarness on systems you own or are authorized to test. Workflows declare risk; scope gates, rate limits, and a circuit breaker constrain execution.

Understand the safety model
passive

Read-only discovery or local analysis.

active-safe

Rate-limited target interaction.

intrusive

Requires a non-empty scope file.

destructive-manual

Requires human approval.

04 / START LOCAL

Your next investigation
starts here.

Install the open-source toolkit.
Keep the traces and evidence on your machine.

Install BountyHarness Inspect the source ↗
curl -fsSL https://raw.githubusercontent.com/Mr-Neutr0n/bounty-harness/main/install.sh | sh -s --