Bug bounty glossary
This bug bounty glossary defines the platform and AI terms used across this site, one line each. Platform definitions are quoted from the platform's own docs.
- AI slop
- Unverified, AI-generated reports; YesWeHack lists "Program spamming and AI Slop" at severity 7. More
- Asset
- A host, app or API the program lists, in or out of scope. More
- Base URL
- The address a tool sends model requests to; changing it changes the provider. More
- Bug bounty program
- A program that pays for valid findings. HackerOne: "In return, the finders of the vulnerabilities are rewarded with monetary prizes." Source
- Context window
- How many tokens a model reads at once; 1,048,576 holds a full policy plus notes. More
- CVSS
- FIRST: "The Common Vulnerability Scoring System (CVSS) provides a way to capture the principal characteristics of a vulnerability and produce a numerical score reflecting its severity." Source
- Duplicate
- HackerOne: "This issue has already been reported or is otherwise previously known by the customer team." Source
- Hackbot
- An automated hunting agent. HackerOne: "Hackbots must not operate in a fully autonomous manner." More
- Impact statement
- Who can do what to whom, under which conditions, with what result. More
- Informative
- HackerOne: "The report contains valid information, but the information provided doesn't require action." Source
- Local model
- A model on your own machine, so notes never leave it. More
- N/A
- HackerOne: "The report doesn't contain a valid reproducible issue, and the security implications have not been demonstrated." Source
- Not Applicable
- Bugcrowd: "A submission that does not apply to the target or application." Source
- OpenAI-compatible
- An API that takes the same Chat Completions requests as OpenAI's, so most tools connect to it. More
- Out of Scope
- Bugcrowd: "A submission which is rejected because it is not in scope with the criteria outlined in the bounty brief." Source
- Proof of concept
- The steps or evidence that show the bug is real and reproducible. More
- Safe harbor
- HackerOne: "Gold Standard Safe Harbor supports the protection of organizations and hackers engaged in Good Faith Security Research." Source
- Scope
- The assets and test types a program allows, and the ones it excludes. More
- Severity
- How serious a finding is, rated with CVSS on HackerOne or the VRT on Bugcrowd. More
- Steps to reproduce
- Numbered steps a triager follows to see the bug. More
- Tokens
- The units models read and bill by; prices here are per 1M tokens. More
- Triage
- The platform or program team checking whether a report is valid, in scope and new. More
- VDP
- HackerOne: "A vulnerability disclosure program solely gives clear guidelines for how an organization would like to be notified of potential security vulnerabilities found by external third parties." Source
- VRT
- Bugcrowd: "Bugcrowd's Vulnerability Rating Taxonomy is a resource outlining Bugcrowd's baseline priority rating, including certain edge cases, for common vulnerabilities." Source