ikooky/velric-agent / skills/velric
velric
VerifiedWeb 应用攻击面测绘与越权检测工具链(velric-agent CLI)。覆盖子域/端口/敏感路径/CORS/WAF 指纹侦察、 SPA/MPA 接口枚举、无头运行时流量录制、前端权限树重建、鉴权面差分(authz-replay/role-matrix/IDOR/mass-assign)、 JWT/GraphQL/Header/Verb 绕过、PII 与凭据泄漏扫描。纯 Node 本地执行,无 LLM 依赖。 触发关键词:攻击面测绘、接口枚举、越权、IDOR、未授权访问、水平垂直越权、参数探测、隐藏参数、 权限树、流量录制、CORS、敏感路径、WAF 指纹、sourcemap 泄漏、JWT 攻击、GraphQL 鉴权、mass assignment、PII 扫描。
velric-agent — Web 攻击面测绘工具链
ACTION REQUIRED(读完后立刻执行)
NOW: 确认目标在授权测试范围内(本工具链仅用于已授权的安全测试)NOW: 校验工具可用 ——bash <repo>/bootstrap.sh --verify通过即就绪;未安装则git clone仓库后bash bootstrap.sh(一条命令装完并自验,之后 skill 直接调用)NEXT: 侦察从references/workflow.md的 Phase 0 开始,工具不是孤立的——有阶段依赖与门禁ACT: 每个命令产物落各自的<outdir>/(JSON+txt 工件),结论必须引用产物文件与真实状态码ACT: 需要出网代理时所有命令共用--proxy URL/--proxy-pool FILE/--rps N全局限速
工具根目录 = velric-agent 仓库路径(记为 $VELRIC,下文命令里替换为 node $VELRIC/scripts/velric-agent.mjs)。
命令矩阵
命令一律形如 node $VELRIC/scripts/velric-agent.mjs <subcommand>,下表省略前缀。首次使用先 bash $VELRIC/bootstrap.sh。命令注册表与分组 = src/cli/registry.ts(CLI help 输出同源渲染)。
侦察 — 资产与面测绘
先跑 fingerprint 判 WAF/类型;subdomains/port-scan/wayback 圈资产;paths/dir-fuzz/info-leak 摸面;cloud-storage/runtime-config-leak 拿桶与运行时配置。
| 命令 | 用途 |
|---|---|
velric-agent subdomains <domain> <outdir> | subdomain enum (crt.sh + CNAME + wordlist) |
velric-agent subdomain-alter <domain> <outdir> --hosts a,b | alteration-based subdomain generation (+ --resolve) |
velric-agent port-scan <host> <outdir> | TCP top-ports scan (--ports/--timeout) |
velric-agent dns-resolve <host> [...] | resolve A/AAAA/CNAME |
velric-agent http-probe <url> | probe one URL (status/title/hash) |
velric-agent enrich <seed...> --out DIR | asset graph from seeds (DNS+HTTP) |
velric-agent wayback <domain> <outdir> | historical URLs from web.archive.org |
velric-agent fingerprint <baseUrl> <outdir> | WAF/CDN id + method enum + cookie audit |
velric-agent paths <baseUrl> <outdir> | sensitive paths w/ soft-404 calibration |
| `velric-agent dir-fuzz (--url URL [--wordlist-file F] | --preset sensitive |
velric-agent nuclei-lite <baseUrl> <outdir> | template matcher (panels/leaks) |
velric-agent admin-surface <baseUrl> <outdir> | admin/panel surface map |
velric-agent cloud-storage <buckets> <outdir> | anonymous S3/GCS/Azure bucket list/read/write (--write) |
velric-agent runtime-config-leak <baseUrl> <outdir> | SPA runtime config leakage (config.js/.env, secret classes) |
velric-agent info-leak <baseUrl> <outdir> | information disclosure (paths/stack/banners/secrets) |
velric-agent coverage <baseUrl> <outdir> | interactive coverage enumeration |
velric-agent site-map <baseUrl> <outdir> | structured site map |
velric-agent asset-graph <dir> <action> | asset graph store |
内容 — 抓取与语料
SPA 走 harvest、MPA 走 spider;响应加密先 envelope-break 破信封;deobfuscate 还原后做后续分析;泄漏语料 git-restore/backup-unpack/sqlite-read。
| 命令 | 用途 |
|---|---|
velric-agent harvest <url> <outdir> | static SPA bundle harvest |
velric-agent spider <url> <outdir> | MPA crawl (--cookie/--max/--depth/--exclude) |
velric-agent extract-routes <jsdir> <outdir> | routes from JS corpus |
velric-agent sourcemaps <jsDir> <outdir> | source-map leak detection (--base URL) |
velric-agent deobfuscate <jsDir> <outdir> | JS corpus deobfuscation (--action analyze) |
velric-agent envelope-break <envelope.json> <outdir> --keys k1,k2 | encrypted response envelope breaker |
velric-agent git-restore <baseUrl> <outdir> | rebuild source from leaked .git |
velric-agent backup-unpack <urls.txt> <outdir> | dump unpack + secret scan |
velric-agent sqlite-read <file.db> <outdir> | SQLite leak reader |
velric-agent archive-recover <url> <outdir> | blocked-page archive recovery |
velric-agent url-refine <outdir> --urls a,b | URL dedupe/classify (--action classify, --fuzzy) |
velric-agent screenshot <outdir> --urls a,b | headless screenshot capture (--cluster) |
velric-agent traffic-recorder <outdir> | traffic recording proxy |
velric-agent traffic-query <dir> <action> | traffic search/get |
运行时 — 浏览器与协议
runtime 抓真实 XHR/WS;merge-params 汇参数矩阵;ws-schema/proto-decode 解私有协议。
| 命令 | 用途 |
|---|---|
velric-agent runtime <url> <outdir> | headless runtime capture (--chromium/--routes/--wait/--per-route/--api-pattern) |
velric-agent session-crawl <baseUrl> <loginUrl> <user> <pass> <outdir> | authenticated crawl w/ relogin |
velric-agent perm-tree <jsdir> <outdir> | permission tree rebuild (--role NAME) |
velric-agent merge-params <samples.json> <outdir> | param matrix merge |
velric-agent ws-schema <runtime_api.json> <outdir> | WS message field schemas |
| `velric-agent proto-decode <hex | base64> ` |
velric-agent ws-origin-check <wsUrl> <outdir> | WS cross-origin check (CSWSH) |
velric-agent hls-chain <gateway-{PATH}> <path> <outdir> --key-url U | HLS capability chain: playlist+key+TS decrypt |
逆向 — 静态与指令层
Electron: asar-unpack 定位 → asar-patch 改写;Flutter: flutter-analyze 挖域/路由/密钥;JS: vm-analyze(自定义 VM opcode)→ dom-sinks/analyze-gates;二进制: elf/wasm/dex-analyze 定性 → wasm/dex/x86-disasm 指令级 + xref。
| 命令 | 用途 |
|---|---|
velric-agent vm-analyze <jsDir> <outdir> | JS VM interpreter static reverse (opcode table, dispatch, state slots) |
velric-agent dom-sinks <jsDir> <outdir> | DOM-XSS taint chain scan |
velric-agent analyze-gates <jsDir> <outdir> | auth-gate static analysis |
velric-agent elf-analyze <binary> <outdir> | ELF static analysis (sections/symbols/hardening/imports) |
velric-agent wasm-analyze <file.wasm> <outdir> | WASM static analysis (imports/exports/symbol recovery) |
| `velric-agent dex-analyze <file.dex | .apk> ` |
velric-agent wasm-disasm <file.wasm> <outdir> | full WASM disassembly (all opcodes, callgraph) |
| `velric-agent dex-disasm <file.dex | .apk> ` |
velric-agent asar-unpack <app.asar> <outdir> | Electron asar extract + license-check localization (--list-only) |
| `velric-agent flutter-analyze <app.apk | libapp.so> ` |
velric-agent asar-patch <app.asar> <out.asar> --patches-file F | asar text patch + repack (crack the license check) |
注入 — 语法与参数
逐差分探针:SQL/NoSQL/LDAP/SSTI 语法面 + 原型污染/HPP/重定向遍历;全部带基线差分与错误签名,findings 需复核语义。
| 命令 | 用途 |
|---|---|
velric-agent sqli-probe <url> <outdir> --param q | SQL injection differential (error/boolean/time blind/auth-bypass) |
velric-agent nosqli-probe <url> <outdir> --param id | NoSQL operator injection ($ne/$gt/$regex/$where, type juggling) |
velric-agent ldap-probe <url> <outdir> --param uid | LDAP injection (wildcard/filter-widening/boolean-OR, row divergence) |
velric-agent ssti-probe <url> <outdir> --param q | SSTI canary arithmetic (11 engine grammars, 49-reflection verdict) |
velric-agent deser-surface <url> <outdir> | deserialization surface (Java/PHP/pickle/YAML, ViewState MAC) |
velric-agent proto-pollute <url> <outdir> | prototype pollution vectors (proto/constructor, JSON body) |
velric-agent hpp-probe <url> <outdir> --params a,b | HTTP parameter pollution (duplicate/array/comma strategies) |
velric-agent redirect-traverse <url> <outdir> --params a,b | open redirect + path traversal encodings (Location analysis, passwd oracle) |
认证 — 凭证与令牌
user-enum 枚举 → spray 定点喷洒 → hash-crack 破 dump 哈希;cred-brute 语义分类;pass-policy/reset-abuse 弱密码与重置接管;jwt-inspect/jwt-attack/jwks-forge 三层 JWT 面;admin-creds 默认口令拿后台;oauth-bypass 重定向劫持;2fa-bypass/reset-deep 第二因子与重置深度;device-login 匿名设备号开号。
| 命令 | 用途 |
|---|---|
velric-agent cred-brute <url> <outdir> --pairs-file F | credential spray (semantic classification) |
| `velric-agent user-enum --mode login | register |
velric-agent spray <url> <outdir> --body-template T --users a,b --passwords p1,p2 | lockout-aware password spray (per-user budget) |
| `velric-agent hash-crack <hash[,hash] | file> ` |
velric-agent otp-brute <url> <outdir> --body-template T | OTP/2FA rate-limit & lockout measurement (+2FA secret leak detection) |
velric-agent pass-policy <url> <outdir> --body-template T | password policy surface (10-tier weak acceptance) |
velric-agent reset-abuse <resetUrl> <loginUrl> <outdir> | static security-answer takeover (spray → reset → login proof) |
velric-agent 2fa-bypass <loginUrl> <verifyUrl> <protectedUrl> <outdir> | 2FA step skip / OTP replay |
velric-agent device-login <loginUrl> <outdir> --cipher-key K | anonymous device-login minting (unlimited accounts) |
velric-agent reset-deep <requestUrl> <outdir> --body-template T --email E | reset token shape/reuse/host-poison/alias depth |
velric-agent jwt-inspect <outdir> | JWT decode + weak-secret crack |
velric-agent jwt-attack <token> <endpoint> <outdir> | JWT attack suite (alg confusion/weak key/kid, --public-key-file) |
velric-agent admin-creds <baseUrl> <outdir> | vendor default-credential takeover on admin panels (10 panels, marker-verified) |
velric-agent oauth-bypass <authzUrl-{REDIRECT}> <legitRedirect> <outdir> | OAuth redirect_uri matrix + PKCE downgrade (takeover) |
velric-agent login-replay <url> <u> <p> <outdir> | login + cookie export |
授权 — 越权与访问控制
auth-probe 画鉴权面 → path-auth-bypass 路径混淆 → signed-url 签名绕过 → idor-prescreen/idor-deep(读写 IDOR)→ authz-replay/role-matrix 多角色差分 → role-escalate/mass-assign/api-version 提权面。
| 命令 | 用途 |
|---|---|
velric-agent auth-probe <api_merged.txt> <outdir> | anonymous vs auth access mapping |
velric-agent path-auth-bypass <protectedUrl> <outdir> | path-confusion filter bypass (20 vectors, soft-404 calibrated) |
velric-agent signed-url <signedUrl> <outdir> | signed-URL enforcement (signature/expiry strip, key case, dot-segment) |
velric-agent idor-prescreen <params_merged.json> <outdir> | IDOR candidate prescreen |
velric-agent idor-deep <outdir> --endpoints[-file] JSON | three-reference IDOR oracle probing (--write for write-IDOR) |
velric-agent authz-replay <recordRoot> <outdir> --roles[-file] JSON | authz differential replay (live-baseline verdicts) |
velric-agent role-matrix <reconRoot> <outdir> --roles[-file] JSON | cross-role access matrix |
velric-agent role-escalate <url> <outdir> --body-template {} | privilege escalation params (role field injection differential) |
velric-agent mass-assign <recordRoot> <outdir> | mass-assignment sink probe (--fields a,b) |
velric-agent api-version <versioned-api-url> <outdir> | API version authz drift (v0..v4 downgrade leaks) |
velric-agent header-bypass <outdir> --urls a,b | auth-header bypass variants (X-Original-URL etc.) |
velric-agent verb-matrix <outdir> --urls a,b | HTTP verb override matrix probe |
velric-agent graphql-authz <endpoint> <schemaFile> <outdir> | GraphQL per-field authz differential |
业务逻辑 — 值域与竞争
value-abuse 负数/溢出注入 + 回读持久化;race-probe TOCTOU 并发自证;cache-deceive/host-poison/upload-probe/graphql-batch 基础设施逻辑面;entitlement-bypass/billing-tamper 订阅与计价绕过。
| 命令 | 用途 |
|---|---|
velric-agent value-abuse <url> <outdir> --body-template T | business-logic value abuse (negative/overflow, read-back persistence) |
velric-agent race-probe <url> <outdir> | TOCTOU race (concurrent mutation + read-back divergence) |
velric-agent cache-deceive <baseUrl> <outdir> --private-path /api/me | web cache deception + unkeyed header poisoning |
velric-agent host-poison <url> <outdir> | Host-header/reset poisoning (9-vector canary reflection) |
velric-agent upload-probe <url> <outdir> | file-upload surface (11 vectors, exec-risk + stored URL) |
velric-agent entitlement-bypass <premiumUrl> <accountUrl> <outdir> | paywall enforcement test (anon/free/claim forgery) |
velric-agent billing-tamper <checkoutUrl> <outdir> --body-template T --amount N | price/currency/qty/coupon tampering with order read-back |
平台 — 管道与证据
auto/auto-batch 确定性管道;hunt 循环台账;waf-adapt/oob-confirm/smuggle/oidc/cors 深化工具;service-unauth/actuator-exploit/actuator-rce 后场直取(未授权设施 / Actuator 挖密钥+Jolokia+RCE);admin-rce 登录后利用;gen-report/save-evidence 收尾。
| 命令 | 用途 |
|---|---|
velric-agent auto <seed> <outdir> | deterministic pipeline (checkpoint resume, --fresh) |
velric-agent auto-batch <seeds.txt> <outdir> | batch pipeline + summary (--notify URL) |
velric-agent run-diff <oldDir> <newDir> | diff two recon runs (attack-surface monitoring) |
velric-agent gen-report <report.json> <outdir> | findings markdown report |
velric-agent save-evidence <items.json> <outdir> | content-addressed evidence blobs |
velric-agent evidence <dir> <action> | evidence blob store |
velric-agent evidence-bind <dir> <action> | evidence binding lifecycle |
velric-agent exploration-graph <dir> <action> | exploration lineage graph |
velric-agent cold-graph <dir> <action> | exploration graph compaction |
velric-agent retest <dir> <action> | finding retest lifecycle |
velric-agent export-findings <f> <dir> <fmt> | findings csv/json/md/sarif export |
velric-agent service-unauth <host> <outdir> | unauthenticated infra sweep (Redis/Docker/etcd/k8s/ES/prom/Nacos...) |
velric-agent actuator-exploit <baseUrl> <outdir> | Spring Actuator: env/heapdump secret mining + gateway SpEL RCE |
velric-agent admin-rce <baseUrl> <panel> <user> <pass> <outdir> | post-login exploit (Tomcat WAR shell/Jenkins console/panel loot) |
velric-agent actuator-rce <baseUrl> <outdir> | Jolokia sysprops/env-write/sessions/heap token extraction |
velric-agent app-api <outdir> --calls-file F | replay app signed-request protocol (HMAC + envelope decrypt) |
velric-agent scan-platform <dir> <action> | native scan platform (projects/tasks/pipeline/assets) |
velric-agent waf-adapt <url> <payload> <outdir> | payload transport adaptation vs WAF (--in query |
velric-agent oob-confirm <url> <outdir> --payloads a,b | OOB callback confirmation (interactsh) |
velric-agent smuggle-probe <baseUrl> <outdir> | request-smuggling differential |
velric-agent smuggle-exploit <httpBaseUrl> <protectedPath> <outdir> | CL.TE/TE.CL smuggling into an admin bypass + h2c upgrade |
velric-agent oidc-probe <baseUrl> <outdir> | OAuth/OIDC/SAML config probe |
velric-agent cors <baseUrl> <outdir> | CORS misconfiguration probe |
velric-agent cors-deep <url> <outdir> | CORS exploitability matrix (reflection × credentials, preflight) |
velric-agent ws-auth <wsUrl> <outdir> | WebSocket auth gap (anonymous upgrade, cross-origin, push; raw RFC6455) |
velric-agent graphql-deep <endpoint> <outdir> | GraphQL introspection + schema |
| `velric-agent mine-params (--url URL | --remove-diff --merged F --base URL)` |
velric-agent probe-params <params_merged.json> <outdir> | per-param differential probe (--max N) |
velric-agent pii-scan <recordRoot> <outdir> | PII/credential exposure scan over recorded traffic |
velric-agent replay-tamper <recordRoot> <outdir> | replay recorded exchanges with tamper rules (--rules-file, --apply) |
hunt 循环(宿主 agent 自主驱动)
| 命令 | 用途 |
|---|---|
velric-agent hunt-begin <goal> <outdir> | 开 hunt(预算 --max-rounds/--max-cost) |
velric-agent hunt-record <outdir> --record-file F | 写回观察(goals/facts/findings/assets;facts/findings 须有台账回执支撑——侦察命令加 --run-id <hunt runId> --observation-ledger-root <outdir> 才产生回执) |
velric-agent hunt-overview <outdir> | 读态势 + 台账证据快照 |
门禁(不可绕过)
- 观察台账反幻觉闸门:叙述的事实/PoC URL 必须来自台账真实记录,状态码一致,否则 drop
- 响应状态码必须真实可见才可写入结论;软 404 页面不得当作命中
- 所有输出到 report 的条目须带证据(文件/URL/状态码)
- 诚实报告:各阶段数量、未覆盖模块、每条结论的来源(静态/运行时/合并)
- secrets/dom-sinks/idor 类预筛结果 = 候选,不是实锤,须复核后才可升级结论
任务完成自检(声称完成前 MUST 通过)
- 每条结论都有产物文件 + 真实状态码支撑(不是"应该存在")
- soft-404 校准后的命中才算命中;401/403 = 存在但受保护,不是漏洞
- 预筛类(secrets/idor/dom-sinks)结论标注"候选",未当实锤
- WAF 在位时已降速降并发(fingerprint 命中 → --rps 全局限速)
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 471k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |