Agent SkillsXiaomingX/mimofan › vuln-hunt

vuln-hunt

GitHub

用于在真实项目中定位并复现漏洞的长周期安全分析技能,涵盖攻击面侦察、假设验证、链追踪及PoC执行。

crates/tui/assets/skills/vuln-hunt/SKILL.md XiaomingX/mimofan

Trigger Scenarios

需要发现并复现具体漏洞 进行深度安全审计

Install

npx skills add XiaomingX/mimofan --skill vuln-hunt -g -y
More Options

Non-standard path

npx skills add https://github.com/XiaomingX/mimofan/tree/main/crates/tui/assets/skills/vuln-hunt -g -y

Use without installing

npx skills use XiaomingX/mimofan@vuln-hunt

指定 Agent (Claude Code)

npx skills add XiaomingX/mimofan --skill vuln-hunt -a claude-code -g -y

安装 repo 全部 skill

npx skills add XiaomingX/mimofan --all -g -y

预览 repo 内 skill

npx skills add XiaomingX/mimofan --list

SKILL.md

Frontmatter
{
    "name": "vuln-hunt",
    "description": "Long-horizon security workflow to FIND AND REPRODUCE a vulnerability in a real project (e.g. fastjson deserialization gadget, Spring Boot RCE) using hypothesis\/evidence tracking, gadget-chain tracing, and a realized run_poc."
}

Vuln Hunt (long-horizon security analysis)

Use this skill when the task is to find AND reproduce a vulnerability in a real project (e.g. fastjson deserialization gadget, Spring Boot RCE), not to merely classify code.

Workflow (follow in order; do not skip steps)

  1. Recon: enumerate attack surface — deps, entry points, sinks (call_graph, ast_query, security_audit/semgrep).
  2. Hypothesize: for each suspected sink, register a hypothesis with hypothesis action=create stating the claimed exploit class + CVE.
  3. Trace: run gadget_chain_trace for the sink; identify which gadgets are PRESENT vs MISSING in the target — the missing ones are what you must prove reachable or rule out. Cross-check with call_graph reachability.
  4. Gather evidence: for every hypothesis, call hypothesis action=add_evidence with concrete code references / taint paths BEFORE concluding.
  5. Reproduce: build a candidate PoC and run run_poc with an expect string that proves the vulnerable behavior. realized=true => exploit confirmed.
  6. Verdict: hypothesis action=resolve verdict=confirmed|refuted — this is REFUSED unless evidence was attached (consistency gate). Never assert a finding without evidence + a realized PoC.

Consistency rules (what gets scored)

  • Every claim must have a registered hypothesis + >=1 evidence entry.
  • A hypothesis with zero evidence CANNOT be resolved.
  • A "confirmed" verdict without a realized run_poc is invalid.
  • Prefer reconstructing the attack PRINCIPLE (why it works) over a bare PoC.

Tooling reference

  • hypothesis: create / add_evidence / resolve / list
  • gadget_chain_trace: sink + present_gadgets -> chain satisfied/missing
  • run_poc: command + expect -> realized bool (executed in sandbox)
  • call_graph / ast_query / security_audit: static traceability

Version History

  • cab98ff Current 2026-08-16 16:04

Same Skill Collection

.workbuddy/skills/parallel-agent-lock-guard/SKILL.md
crates/tui/assets/skills/delegate/SKILL.md
crates/tui/assets/skills/documents/SKILL.md
crates/tui/assets/skills/feishu/SKILL.md
crates/tui/assets/skills/fleet-manager/SKILL.md
crates/tui/assets/skills/mcp-builder/SKILL.md
crates/tui/assets/skills/pdf/SKILL.md
crates/tui/assets/skills/plugin-creator/SKILL.md
crates/tui/assets/skills/presentations/SKILL.md
crates/tui/assets/skills/security-audit/SKILL.md
crates/tui/assets/skills/skill-creator/SKILL.md
crates/tui/assets/skills/skill-installer/SKILL.md
crates/tui/assets/skills/spreadsheets/SKILL.md
crates/tui/assets/skills/v4-best-practices/SKILL.md

Metadata

Files
0
Version
cab98ff
Hash
672b5cef
Indexed
2026-08-16 16:04

Главная - Вики-сайт
Copyright © 2011-2026 iteam. Current version is 2.155.2. UTC+08:00, 2026-08-16 19:34
浙ICP备14020137号-1 $Гость$