AGENT SKILL

bug-bounty-hunter

Run an effective bug bounty program — scope design, triage, researcher relations, and remediation workflows.

License 许可
MIT
所属仓库
TanNhatCMS/jarvis
最近更新
2026-11-07
Stars
★ 0
Forks
1

01 前言 Front Matter

name
bug-bounty-hunter
category
security
description
Run an effective bug bounty program — scope design, triage, researcher relations, and remediation workflows.

03 SKILL.md 正文

Overview

A bug bounty program pays independent researchers for responsibly disclosed vulnerabilities in your systems. Done well, it is continuous, adversarial testing at a scale no internal team can match. Done badly, it is a flood of duplicates, out-of-scope reports, and burned researcher goodwill.

This skill is written for program owners (defenders running the program) — scoping, triage, payouts, and turning reports into fixes. It covers only authorized, in-scope testing methodology at the standard published level: recon, testing within the published scope, and quality reporting. No exploit code.

A bounty program is a marketplace: researchers sell attention and skill, you buy findings with money, responsiveness, and respect. Programs that thrive treat researchers as an extension of the security team — fast triage, fair pay, technical feedback, and public credit where wanted. Programs that treat them as ticket-generators get what they pay for.

When to use

  • Launching a private or public bounty program (start private, invite-only).
  • Triaging incoming reports quickly and fairly.
  • Setting bounty tables, safe harbor language, and scope that attracts good researchers.
  • Integrating bounty findings into your SDLC so reports become fixes, not backlog.
  • Deciding whether your org is mature enough for bounty (you need vuln-management first).

Core concepts

  • Safe harbor: an explicit commitment not to pursue legal action for good-faith, in-scope research. Without it, good researchers stay away.

  • Scope is the product: clear in-scope assets, explicit out-of-scope items, and testing rules. Vague scope produces vague reports and disputes.

  • Private before public: start with a small vetted researcher pool, tune triage, then open up. Public launch with broken triage is reputational damage.

  • Triage SLA: acknowledge in 1–2 days, first assessment in under a week. Slow triage kills researcher motivation faster than low payouts.

  • Duplicates are data: high duplicate rates on an asset mean your vuln-management is behind, not that researchers are sloppy. Fix the root cause.

  • The report is the deliverable: severity, reproduction steps, impact in your context, suggested fix. Reward quality reports, not just criticals.

  • Signal over volume. Ten high-quality researchers outperform a thousand drive-by submitters. Invest in relationships with proven researchers: private scopes, bonuses, early access.

  • Root-cause bounties. Consider bonus multipliers for reports that identify systemic issues or come with suggested fixes — it shifts incentives from bugs to classes.

  • Disclosure policy. A clear coordinated-disclosure policy with timelines builds researcher trust and turns fixes into positive publicity.

Practical workflow

  1. Readiness check: do you have asset inventory, a vuln-management process with SLAs, and someone to own triage? If not, fix that first — bounty amplifies process gaps.
  2. Design the program: define scope (start narrow: one app or API), rules of engagement, out-of-scope list, safe harbor text, bounty table by severity, and response SLAs. Publish it all.
  3. Launch private: invite 20–50 researchers, run 4–8 weeks, tune scope and triage on real volume.
  4. Triage pipeline: intake → deduplicate → validate (reproduce in-house) → severity using your rubric → route to the owning team with a fix SLA → verify the fix → pay → disclose (if policy allows).
  5. Researcher relations: communicate decisions with reasons, pay promptly, give feedback on report quality, invite strong researchers to private scopes. Reputation compounds.
  6. Close the loop: track bounty findings in the same system as pentest/scanner findings; measure time-to-fix and recurrence. Feed common bug classes into secure-coding training and SAST rules.
  7. Go public when ready: triage holding up, fix SLAs holding up, scope stable. Announce with a clear brief and realistic expectations.

Bounty brief essentials

  • In-scope assets (exact domains, apps, APIs, mobile builds)
  • Out-of-scope (third-party services, DoS, spam, social engineering unless invited)
  • Rules: no data exfiltration beyond proof, no degrading service, stop at PII
  • Safe harbor statement
  • Severity rubric and payout table
  • Response and payout SLAs

Sustaining the practice

  • Review the bounty table annually against market rates and program data
  • Publish anonymized program stats to attract quality researchers
  • Rotate fresh private scopes to retain top researchers
  • Audit triage quality by sampling closed reports quarterly

Metrics that prove it works

  • Median time to first triage response and to severity decision
  • Time from validation to fix, by severity
  • Valid-report rate (signal quality of the program)
  • Researcher retention: % of top researchers still active after 6 months

Common pitfalls

  • Launching public on day one. You will drown. Private first, always.
  • Scope ambiguity. "Our main site" is not scope. Enumerate assets explicitly.
  • Slow or silent triage. Researchers talk. A program known for ghosting gets low-effort reports.
  • Paying late or disputing fairly-earned bounties. Nothing kills a program faster.
  • Treating bounty as your vuln-management. It is a complement to scanning, pentests, and code review — not a replacement.
  • Ignoring duplicates as signal. Ten duplicate IDOR reports = your access-control testing is broken. Fix the class.
  • No fix verification. Paying without verifying the fix teaches researchers (and your team) that closure does not matter.
  • Scope that excludes everything interesting. Researchers invest where scope rewards them. Overly narrow scope gets low-effort reports on the allowed sliver.
  • Inconsistent severity calls. Publish your rubric with worked examples. Nothing erodes trust like two identical bugs paid differently.
  • Scope creep by researcher pressure. Expanding scope reactively to appease complaints creates inconsistency. Change scope deliberately, announce clearly.
  • Ignoring the long tail of low-severity reports. Individually minor, in aggregate they map your attack surface. Mine them for patterns quarterly.

04 同仓库更多 Skills

TanNhatCMS
TanNhatCMS/jarvis

web-scrape

Turn web pages into LLM-ready markdown via the Firecrawl API (renders JS, anti-bot handling). Use when browser_open/fetch fails on JS-heavy pages, or when you need clean markdown/structured extraction from a site. Requires FIRECRAWL_API_KEY.

★ 0
TanNhatCMS
TanNhatCMS/jarvis

design-taste-frontend

tasteskill v2, the anti-slop frontend skill. Art direction for websites, distilled from a study of 247 curated gallery sites (One Page Love, Awwwards Sites of the Day). Deals researched directions (idea, composition, type, palette, photographs, section blueprints) from a library, enforces a strict no-slop house style with a real-browser inspector, and runs a look-and-fix loop on screenshots. Use this whenever you build, design, redesign, restyle, polish or review a landing page, marketing site, portfolio, product page, one-pager, hero or any public-facing web page, in any stack (HTML/CSS, React, Next.js, Vue, Svelte, Astro, Tailwind) — also when the user says a page looks generic, templated, "AI-made", sloppy, or asks to make it look designed, premium, award-level or less like every other site. Not for dashboards or app screens.

★ 0
TanNhatCMS
TanNhatCMS/jarvis

test-driven-development

Use when implementing any feature or bugfix, before writing implementation code

★ 0
TanNhatCMS
TanNhatCMS/jarvis

systematic-debugging

Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes

★ 0
TanNhatCMS
TanNhatCMS/jarvis

requesting-code-review

Use when completing tasks, implementing major features, or before merging to verify work meets requirements

★ 0
TanNhatCMS
TanNhatCMS/jarvis

brainstorming

You MUST use this before any creative work - creating features, building components, adding functionality, modifying behavior, or planning anything new, in software or out of it (a talk, a business, a renovation).

★ 0