main.go 23 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672673674675676677678679680681682683684685686687688689690691692693694695696697698699700701702703704705706707708709710711712713714715716717718719720721722723724725726727728729730731732733734735736737738739740741742743744745746747748749750751752753754755756757758759760761762763764765766767768769770771772773774775776777778779780781782783784785786787788789790791792793794795796797798799800801802803804805806807808809810811812813814815816817818819820821822823824825826827828829830831832833834835836837838839840841842843844845846847848849850851852853854855856857858859860861862863864865866867868869870871872873
  1. package main
  2. import (
  3. "crypto/md5"
  4. "encoding/csv"
  5. "encoding/json"
  6. "errors"
  7. "fmt"
  8. "io/ioutil"
  9. "math"
  10. "net"
  11. "net/url"
  12. "os"
  13. "os/user"
  14. "path"
  15. "path/filepath"
  16. "regexp"
  17. "runtime"
  18. "strconv"
  19. "strings"
  20. "sync"
  21. "time"
  22. "gopkg.in/src-d/go-git.v4/plumbing"
  23. "gopkg.in/src-d/go-git.v4/plumbing/object"
  24. "gopkg.in/src-d/go-git.v4/plumbing/transport/ssh"
  25. "gopkg.in/src-d/go-git.v4/storage/memory"
  26. "github.com/BurntSushi/toml"
  27. "github.com/google/go-github/github"
  28. "github.com/hako/durafmt"
  29. flags "github.com/jessevdk/go-flags"
  30. log "github.com/sirupsen/logrus"
  31. git "gopkg.in/src-d/go-git.v4"
  32. )
  33. // Leak represents a leaked secret or regex match.
  34. type Leak struct {
  35. Line string `json:"line"`
  36. Commit string `json:"commit"`
  37. Offender string `json:"offender"`
  38. Type string `json:"reason"`
  39. Message string `json:"commitMsg"`
  40. Author string `json:"author"`
  41. File string `json:"file"`
  42. Repo string `json:"repo"`
  43. Date time.Time `json:"date"`
  44. }
  45. // RepoDescriptor contains a src-d git repository and other data about the repo
  46. type RepoDescriptor struct {
  47. path string
  48. url string
  49. name string
  50. repository *git.Repository
  51. err error
  52. }
  53. // Options for gitleaks
  54. type Options struct {
  55. // remote target options
  56. Repo string `short:"r" long:"repo" description:"Repo url to audit"`
  57. GithubUser string `long:"github-user" description:"Github user to audit"`
  58. GithubOrg string `long:"github-org" description:"Github organization to audit"`
  59. GithubURL string `long:"github-url" default:"https://api.github.com/" description:"GitHub API Base URL, use for GitHub Enterprise. Example: https://github.example.com/api/v3/"`
  60. GithubPR string `long:"github-pr" description:"Github PR url to audit. This does not clone the repo. GITHUB_TOKEN must be set"`
  61. /*
  62. TODO:
  63. GitLabUser string `long:"gitlab-user" description:"User url to audit"`
  64. GitLabOrg string `long:"gitlab-org" description:"Organization url to audit"`
  65. */
  66. Commit string `short:"c" long:"commit" description:"sha of commit to stop at"`
  67. Depth int `long:"depth" description:"maximum commit depth"`
  68. // local target option
  69. RepoPath string `long:"repo-path" description:"Path to repo"`
  70. OwnerPath string `long:"owner-path" description:"Path to owner directory (repos discovered)"`
  71. // Process options
  72. Threads int `long:"threads" description:"Maximum number of threads gitleaks spawns"`
  73. Disk bool `long:"disk" description:"Clones repo(s) to disk"`
  74. SingleSearch string `long:"single-search" description:"single regular expression to search for"`
  75. ConfigPath string `long:"config" description:"path to gitleaks config"`
  76. SSHKey string `long:"ssh-key" description:"path to ssh key"`
  77. ExcludeForks bool `long:"exclude-forks" description:"exclude forks for organization/user audits"`
  78. Entropy float64 `long:"entropy" short:"e" description:"Include entropy checks during audit. Entropy scale: 0.0(no entropy) - 8.0(max entropy)"`
  79. // TODO: IncludeMessages string `long:"messages" description:"include commit messages in audit"`
  80. // Output options
  81. Log string `short:"l" long:"log" description:"log level"`
  82. Verbose bool `short:"v" long:"verbose" description:"Show verbose output from gitleaks audit"`
  83. Report string `long:"report" description:"path to write report file. Needs to be csv or json"`
  84. Redact bool `long:"redact" description:"redact secrets from log messages and report"`
  85. Version bool `long:"version" description:"version number"`
  86. SampleConfig bool `long:"sample-config" description:"prints a sample config file"`
  87. }
  88. // Config struct for regexes matching and whitelisting
  89. type Config struct {
  90. Regexes []struct {
  91. Description string
  92. Regex string
  93. }
  94. Whitelist struct {
  95. Files []string
  96. Regexes []string
  97. Commits []string
  98. Repos []string
  99. }
  100. Misc struct {
  101. Entropy []string
  102. }
  103. }
  104. type gitDiff struct {
  105. content string
  106. commit *object.Commit
  107. filePath string
  108. repoName string
  109. githubCommit *github.RepositoryCommit
  110. sha string
  111. message string
  112. author string
  113. date time.Time
  114. }
  115. type entropyRange struct {
  116. v1 float64
  117. v2 float64
  118. }
  119. const defaultGithubURL = "https://api.github.com/"
  120. const version = "1.19.0"
  121. const errExit = 2
  122. const leakExit = 1
  123. const defaultConfig = `
  124. # This is a sample config file for gitleaks. You can configure gitleaks what to search for and what to whitelist.
  125. # The output you are seeing here is the default gitleaks config. If GITLEAKS_CONFIG environment variable
  126. # is set, gitleaks will load configurations from that path. If option --config-path is set, gitleaks will load
  127. # configurations from that path. Gitleaks does not whitelist anything by default.
  128. title = "gitleaks config"
  129. # add regexes to the regex table
  130. [[regexes]]
  131. description = "AWS"
  132. regex = '''AKIA[0-9A-Z]{16}'''
  133. [[regexes]]
  134. description = "PKCS8"
  135. regex = '''-----BEGIN PRIVATE KEY-----'''
  136. [[regexes]]
  137. description = "RSA"
  138. regex = '''-----BEGIN RSA PRIVATE KEY-----'''
  139. [[regexes]]
  140. description = "SSH"
  141. regex = '''-----BEGIN OPENSSH PRIVATE KEY-----'''
  142. [[regexes]]
  143. description = "PGP"
  144. regex = '''-----BEGIN PGP PRIVATE KEY BLOCK-----'''
  145. [[regexes]]
  146. description = "Facebook"
  147. regex = '''(?i)facebook(.{0,4})?['\"][0-9a-f]{32}['\"]'''
  148. [[regexes]]
  149. description = "Twitter"
  150. regex = '''(?i)twitter(.{0,4})?['\"][0-9a-zA-Z]{35,44}['\"]'''
  151. [[regexes]]
  152. description = "Github"
  153. regex = '''(?i)github(.{0,4})?['\"][0-9a-zA-Z]{35,40}['\"]'''
  154. [[regexes]]
  155. description = "Slack"
  156. regex = '''xox[baprs]-([0-9a-zA-Z]{10,48})?'''
  157. [whitelist]
  158. #commits = [
  159. # "BADHA5H1",
  160. # "BADHA5H2",
  161. #]
  162. #repos = [
  163. # "mygoodrepo"
  164. #]
  165. [misc]
  166. #entropy = [
  167. # "3.3-4.30"
  168. # "6.0-8.0
  169. #]
  170. `
  171. var (
  172. opts Options
  173. regexes map[string]*regexp.Regexp
  174. singleSearchRegex *regexp.Regexp
  175. whiteListRegexes []*regexp.Regexp
  176. whiteListFiles []*regexp.Regexp
  177. whiteListCommits map[string]bool
  178. whiteListRepos []*regexp.Regexp
  179. entropyRanges []entropyRange
  180. fileDiffRegex *regexp.Regexp
  181. sshAuth *ssh.PublicKeys
  182. dir string
  183. threads int
  184. totalCommits int64
  185. commitMap = make(map[string]bool)
  186. cMutex = &sync.Mutex{}
  187. )
  188. func init() {
  189. log.SetOutput(os.Stdout)
  190. // threads = runtime.GOMAXPROCS(0) / 2
  191. threads = 1
  192. regexes = make(map[string]*regexp.Regexp)
  193. whiteListCommits = make(map[string]bool)
  194. }
  195. func main() {
  196. parser := flags.NewParser(&opts, flags.Default)
  197. _, err := parser.Parse()
  198. if err != nil {
  199. if flagsErr, ok := err.(*flags.Error); ok && flagsErr.Type == flags.ErrHelp {
  200. os.Exit(0)
  201. }
  202. }
  203. if len(os.Args) == 1 {
  204. parser.WriteHelp(os.Stdout)
  205. os.Exit(0)
  206. }
  207. if opts.Version {
  208. fmt.Println(version)
  209. os.Exit(0)
  210. }
  211. if opts.SampleConfig {
  212. fmt.Println(defaultConfig)
  213. os.Exit(0)
  214. }
  215. now := time.Now()
  216. leaks, err := run()
  217. if err != nil {
  218. if strings.Contains(err.Error(), "whitelisted") {
  219. log.Info(err.Error())
  220. os.Exit(0)
  221. }
  222. log.Error(err)
  223. os.Exit(errExit)
  224. }
  225. if opts.Report != "" {
  226. writeReport(leaks)
  227. }
  228. if len(leaks) != 0 {
  229. log.Warnf("%d leaks detected. %d commits inspected in %s", len(leaks), totalCommits, durafmt.Parse(time.Now().Sub(now)).String())
  230. os.Exit(leakExit)
  231. } else {
  232. log.Infof("%d leaks detected. %d commits inspected in %s", len(leaks), totalCommits, durafmt.Parse(time.Now().Sub(now)).String())
  233. }
  234. }
  235. // run parses options and kicks off the audit
  236. func run() ([]Leak, error) {
  237. var leaks []Leak
  238. setLogs()
  239. err := optsGuard()
  240. if err != nil {
  241. return nil, err
  242. }
  243. err = loadToml()
  244. if err != nil {
  245. return nil, err
  246. }
  247. sshAuth, err = getSSHAuth()
  248. if err != nil {
  249. return leaks, err
  250. }
  251. if opts.Disk {
  252. // temporary directory where all the gitleaks plain clones will reside
  253. dir, err = ioutil.TempDir("", "gitleaks")
  254. defer os.RemoveAll(dir)
  255. if err != nil {
  256. return nil, err
  257. }
  258. }
  259. // start audits
  260. if opts.Repo != "" || opts.RepoPath != "" {
  261. // Audit a single remote repo or a local repo.
  262. repo, err := cloneRepo()
  263. if err != nil {
  264. return leaks, err
  265. }
  266. return auditGitRepo(repo)
  267. } else if opts.OwnerPath != "" {
  268. // Audit local repos. Gitleaks will look for all child directories of OwnerPath for
  269. // git repos and perform an audit on said repos.
  270. repos, err := discoverRepos(opts.OwnerPath)
  271. if err != nil {
  272. return leaks, err
  273. }
  274. for _, repo := range repos {
  275. leaksFromRepo, err := auditGitRepo(repo)
  276. if err != nil {
  277. return leaks, err
  278. }
  279. leaks = append(leaksFromRepo, leaks...)
  280. }
  281. } else if opts.GithubOrg != "" || opts.GithubUser != "" {
  282. // Audit a github owner -- a user or organization.
  283. leaks, err = auditGithubRepos()
  284. if err != nil {
  285. return leaks, err
  286. }
  287. } else if opts.GithubPR != "" {
  288. return auditGithubPR()
  289. }
  290. return leaks, nil
  291. }
  292. // writeReport writes a report to a file specified in the --report= option.
  293. // Default format for report is JSON. You can use the --csv option to write the report as a csv
  294. func writeReport(leaks []Leak) error {
  295. var err error
  296. log.Infof("writing report to %s", opts.Report)
  297. if strings.HasSuffix(opts.Report, ".csv") {
  298. f, err := os.Create(opts.Report)
  299. if err != nil {
  300. return err
  301. }
  302. defer f.Close()
  303. w := csv.NewWriter(f)
  304. w.Write([]string{"repo", "line", "commit", "offender", "reason", "commitMsg", "author", "file", "date"})
  305. for _, leak := range leaks {
  306. w.Write([]string{leak.Repo, leak.Line, leak.Commit, leak.Offender, leak.Type, leak.Message, leak.Author, leak.File, leak.Date.Format(time.RFC3339)})
  307. }
  308. w.Flush()
  309. } else {
  310. reportJSON, _ := json.MarshalIndent(leaks, "", "\t")
  311. err = ioutil.WriteFile(opts.Report, reportJSON, 0644)
  312. }
  313. return err
  314. }
  315. // cloneRepo clones a repo to memory(default) or to disk if the --disk option is set.
  316. func cloneRepo() (*RepoDescriptor, error) {
  317. var (
  318. err error
  319. repo *git.Repository
  320. )
  321. // check if whitelist
  322. for _, re := range whiteListRepos {
  323. if re.FindString(opts.Repo) != "" {
  324. return nil, fmt.Errorf("skipping %s, whitelisted", opts.Repo)
  325. }
  326. }
  327. if opts.Disk {
  328. log.Infof("cloning %s", opts.Repo)
  329. cloneTarget := fmt.Sprintf("%s/%x", dir, md5.Sum([]byte(fmt.Sprintf("%s%s", opts.GithubUser, opts.Repo))))
  330. if strings.HasPrefix(opts.Repo, "git") {
  331. repo, err = git.PlainClone(cloneTarget, false, &git.CloneOptions{
  332. URL: opts.Repo,
  333. Progress: os.Stdout,
  334. Auth: sshAuth,
  335. })
  336. } else {
  337. repo, err = git.PlainClone(cloneTarget, false, &git.CloneOptions{
  338. URL: opts.Repo,
  339. Progress: os.Stdout,
  340. })
  341. }
  342. } else if opts.RepoPath != "" {
  343. log.Infof("opening %s", opts.RepoPath)
  344. repo, err = git.PlainOpen(opts.RepoPath)
  345. } else {
  346. log.Infof("cloning %s", opts.Repo)
  347. if strings.HasPrefix(opts.Repo, "git") {
  348. repo, err = git.Clone(memory.NewStorage(), nil, &git.CloneOptions{
  349. URL: opts.Repo,
  350. Progress: os.Stdout,
  351. Auth: sshAuth,
  352. })
  353. } else {
  354. repo, err = git.Clone(memory.NewStorage(), nil, &git.CloneOptions{
  355. URL: opts.Repo,
  356. Progress: os.Stdout,
  357. })
  358. }
  359. }
  360. return &RepoDescriptor{
  361. repository: repo,
  362. path: opts.RepoPath,
  363. url: opts.Repo,
  364. name: filepath.Base(opts.Repo),
  365. err: err,
  366. }, nil
  367. }
  368. // auditGitRepo beings an audit on a git repository
  369. func auditGitRepo(repo *RepoDescriptor) ([]Leak, error) {
  370. var (
  371. err error
  372. leaks []Leak
  373. )
  374. for _, re := range whiteListRepos {
  375. if re.FindString(repo.name) != "" {
  376. return leaks, fmt.Errorf("skipping %s, whitelisted", repo.name)
  377. }
  378. }
  379. // clear commit cache
  380. commitMap = make(map[string]bool)
  381. refs, err := repo.repository.Storer.IterReferences()
  382. if err != nil {
  383. return leaks, err
  384. }
  385. err = refs.ForEach(func(ref *plumbing.Reference) error {
  386. if ref.Name().IsTag() {
  387. return nil
  388. }
  389. branchLeaks := auditGitReference(repo, ref)
  390. for _, leak := range branchLeaks {
  391. leaks = append(leaks, leak)
  392. }
  393. return nil
  394. })
  395. return leaks, err
  396. }
  397. // auditGitReference beings the audit for a git reference. This function will
  398. // traverse the git reference and audit each line of each diff.
  399. func auditGitReference(repo *RepoDescriptor, ref *plumbing.Reference) []Leak {
  400. var (
  401. err error
  402. repoName string
  403. leaks []Leak
  404. commitCount int
  405. commitWg sync.WaitGroup
  406. mutex = &sync.Mutex{}
  407. semaphore chan bool
  408. )
  409. repoName = repo.name
  410. if opts.Threads != 0 {
  411. threads = opts.Threads
  412. }
  413. if opts.RepoPath != "" {
  414. threads = 1
  415. }
  416. semaphore = make(chan bool, threads)
  417. cIter, err := repo.repository.Log(&git.LogOptions{From: ref.Hash()})
  418. if err != nil {
  419. return nil
  420. }
  421. err = cIter.ForEach(func(c *object.Commit) error {
  422. if commitMap[c.Hash.String()] {
  423. return nil
  424. }
  425. cMutex.Lock()
  426. commitMap[c.Hash.String()] = true
  427. cMutex.Unlock()
  428. if c == nil || c.Hash.String() == opts.Commit || (opts.Depth != 0 && commitCount == opts.Depth) {
  429. cIter.Close()
  430. return errors.New("ErrStop")
  431. }
  432. commitCount = commitCount + 1
  433. totalCommits = totalCommits + 1
  434. if whiteListCommits[c.Hash.String()] {
  435. log.Infof("skipping commit: %s\n", c.Hash.String())
  436. return nil
  437. }
  438. err = c.Parents().ForEach(func(parent *object.Commit) error {
  439. commitWg.Add(1)
  440. semaphore <- true
  441. go func(c *object.Commit, parent *object.Commit) {
  442. var (
  443. filePath string
  444. skipFile bool
  445. )
  446. defer func() {
  447. commitWg.Done()
  448. <-semaphore
  449. if r := recover(); r != nil {
  450. log.Warnf("recoverying from panic on commit %s, likely large diff causing panic", c.Hash.String())
  451. }
  452. }()
  453. patch, err := c.Patch(parent)
  454. if err != nil {
  455. log.Warnf("problem generating patch for commit: %s\n", c.Hash.String())
  456. return
  457. }
  458. for _, f := range patch.FilePatches() {
  459. if f.IsBinary() {
  460. continue
  461. }
  462. skipFile = false
  463. from, to := f.Files()
  464. filePath = "???"
  465. if from != nil {
  466. filePath = from.Path()
  467. } else if to != nil {
  468. filePath = to.Path()
  469. }
  470. for _, re := range whiteListFiles {
  471. if re.FindString(filePath) != "" {
  472. skipFile = true
  473. break
  474. }
  475. }
  476. if skipFile {
  477. continue
  478. }
  479. chunks := f.Chunks()
  480. for _, chunk := range chunks {
  481. if chunk.Type() == 1 || chunk.Type() == 2 {
  482. diff := gitDiff{
  483. repoName: repoName,
  484. filePath: filePath,
  485. content: chunk.Content(),
  486. sha: c.Hash.String(),
  487. author: c.Author.String(),
  488. message: c.Message,
  489. date: c.Author.When,
  490. }
  491. chunkLeaks := inspect(diff)
  492. for _, leak := range chunkLeaks {
  493. mutex.Lock()
  494. leaks = append(leaks, leak)
  495. mutex.Unlock()
  496. }
  497. }
  498. }
  499. }
  500. }(c, parent)
  501. return nil
  502. })
  503. return nil
  504. })
  505. commitWg.Wait()
  506. return leaks
  507. }
  508. // inspect will parse each line of the git diff's content against a set of regexes or
  509. // a set of regexes set by the config (see gitleaks.toml for example). This function
  510. // will skip lines that include a whitelisted regex. A list of leaks is returned.
  511. // If verbose mode (-v/--verbose) is set, then checkDiff will log leaks as they are discovered.
  512. func inspect(diff gitDiff) []Leak {
  513. lines := strings.Split(diff.content, "\n")
  514. var (
  515. leaks []Leak
  516. skipLine bool
  517. )
  518. for _, line := range lines {
  519. skipLine = false
  520. for leakType, re := range regexes {
  521. match := re.FindString(line)
  522. if match == "" {
  523. continue
  524. }
  525. // if offender matches whitelist regex, ignore it
  526. for _, wRe := range whiteListRegexes {
  527. whitelistMatch := wRe.FindString(line)
  528. if whitelistMatch != "" {
  529. skipLine = true
  530. break
  531. }
  532. }
  533. if skipLine {
  534. break
  535. }
  536. leaks = addLeak(leaks, line, match, leakType, diff)
  537. }
  538. if opts.Entropy > 0 || len(entropyRanges) != 0 {
  539. entropyLeak := false
  540. words := strings.Fields(line)
  541. for _, word := range words {
  542. entropy := getShannonEntropy(word)
  543. if entropy >= opts.Entropy && len(entropyRanges) == 0 {
  544. entropyLeak = true
  545. }
  546. if len(entropyRanges) != 0 {
  547. for _, eR := range entropyRanges {
  548. if entropy > eR.v1 && entropy < eR.v2 {
  549. entropyLeak = true
  550. }
  551. }
  552. }
  553. if entropyLeak {
  554. leaks = addLeak(leaks, line, word, fmt.Sprintf("Entropy: %.2f", entropy), diff)
  555. entropyLeak = false
  556. }
  557. }
  558. }
  559. }
  560. return leaks
  561. }
  562. // addLeak is helper for func inspect() to append leaks if found during a diff check.
  563. func addLeak(leaks []Leak, line string, offender string, leakType string, diff gitDiff) []Leak {
  564. leak := Leak{
  565. Line: line,
  566. Commit: diff.sha,
  567. Offender: offender,
  568. Type: leakType,
  569. Author: diff.author,
  570. File: diff.filePath,
  571. Repo: diff.repoName,
  572. Message: diff.message,
  573. Date: diff.date,
  574. }
  575. if opts.Redact {
  576. leak.Offender = "REDACTED"
  577. leak.Line = "REDACTED"
  578. }
  579. if opts.Verbose {
  580. leak.log()
  581. }
  582. leaks = append(leaks, leak)
  583. return leaks
  584. }
  585. // getShannonEntropy https://en.wiktionary.org/wiki/Shannon_entropy
  586. func getShannonEntropy(data string) (entropy float64) {
  587. if data == "" {
  588. return 0
  589. }
  590. charCounts := make(map[rune]int)
  591. for _, char := range data {
  592. charCounts[char]++
  593. }
  594. invLength := 1.0 / float64(len(data))
  595. for _, count := range charCounts {
  596. freq := float64(count) * invLength
  597. entropy -= freq * math.Log2(freq)
  598. }
  599. return entropy
  600. }
  601. // discoverRepos walks all the children of `path`. If a child directory
  602. // contain a .git file then that repo will be added to the list of repos returned
  603. func discoverRepos(ownerPath string) ([]*RepoDescriptor, error) {
  604. var (
  605. err error
  606. repos []*RepoDescriptor
  607. )
  608. files, err := ioutil.ReadDir(ownerPath)
  609. if err != nil {
  610. return repos, err
  611. }
  612. for _, f := range files {
  613. if f.IsDir() {
  614. repoPath := path.Join(ownerPath, f.Name())
  615. r, err := git.PlainOpen(repoPath)
  616. if err != nil {
  617. continue
  618. }
  619. repos = append(repos, &RepoDescriptor{
  620. repository: r,
  621. name: f.Name(),
  622. path: repoPath,
  623. })
  624. }
  625. }
  626. return repos, err
  627. }
  628. // setLogLevel sets log level for gitleaks. Default is Warning
  629. func setLogs() {
  630. switch opts.Log {
  631. case "info":
  632. log.SetLevel(log.InfoLevel)
  633. case "debug":
  634. log.SetLevel(log.DebugLevel)
  635. case "warn":
  636. log.SetLevel(log.WarnLevel)
  637. default:
  638. log.SetLevel(log.InfoLevel)
  639. }
  640. log.SetFormatter(&log.TextFormatter{
  641. FullTimestamp: true,
  642. })
  643. }
  644. // optsGuard prevents invalid options
  645. func optsGuard() error {
  646. var err error
  647. if opts.GithubOrg != "" && opts.GithubUser != "" {
  648. return fmt.Errorf("github user and organization set")
  649. } else if opts.GithubOrg != "" && opts.OwnerPath != "" {
  650. return fmt.Errorf("github organization set and local owner path")
  651. } else if opts.GithubUser != "" && opts.OwnerPath != "" {
  652. return fmt.Errorf("github user set and local owner path")
  653. }
  654. if opts.Threads > runtime.GOMAXPROCS(0) {
  655. return fmt.Errorf("%d available threads", runtime.GOMAXPROCS(0))
  656. }
  657. // do the URL Parse and error checking here, so we can skip it later
  658. // empty string is OK, it will default to the public github URL.
  659. if opts.GithubURL != "" && opts.GithubURL != defaultGithubURL {
  660. if !strings.HasSuffix(opts.GithubURL, "/") {
  661. opts.GithubURL += "/"
  662. }
  663. ghURL, err := url.Parse(opts.GithubURL)
  664. if err != nil {
  665. return err
  666. }
  667. tcpPort := "443"
  668. if ghURL.Scheme == "http" {
  669. tcpPort = "80"
  670. }
  671. timeout := time.Duration(1 * time.Second)
  672. _, err = net.DialTimeout("tcp", ghURL.Host+":"+tcpPort, timeout)
  673. if err != nil {
  674. return fmt.Errorf("%s unreachable, error: %s", ghURL.Host, err)
  675. }
  676. }
  677. if opts.SingleSearch != "" {
  678. singleSearchRegex, err = regexp.Compile(opts.SingleSearch)
  679. if err != nil {
  680. return fmt.Errorf("unable to compile regex: %s, %v", opts.SingleSearch, err)
  681. }
  682. }
  683. if opts.Entropy > 8 {
  684. return fmt.Errorf("The maximum level of entropy is 8")
  685. }
  686. if opts.Report != "" {
  687. if !strings.HasSuffix(opts.Report, ".json") && !strings.HasSuffix(opts.Report, ".csv") {
  688. return fmt.Errorf("Report should be a .json or .csv file")
  689. }
  690. dirPath := filepath.Dir(opts.Report)
  691. if _, err := os.Stat(dirPath); os.IsNotExist(err) {
  692. return fmt.Errorf("%s does not exist", dirPath)
  693. }
  694. }
  695. return nil
  696. }
  697. // loadToml loads of the toml config containing regexes and whitelists.
  698. // This function will first look if the configPath is set and load the config
  699. // from that file. Otherwise will then look for the path set by the GITHLEAKS_CONIFG
  700. // env var. If that is not set, then gitleaks will continue with the default configs
  701. // specified by the const var at the top `defaultConfig`
  702. func loadToml() error {
  703. var (
  704. config Config
  705. configPath string
  706. )
  707. if opts.ConfigPath != "" {
  708. configPath = opts.ConfigPath
  709. _, err := os.Stat(configPath)
  710. if err != nil {
  711. return fmt.Errorf("no gitleaks config at %s", configPath)
  712. }
  713. } else {
  714. configPath = os.Getenv("GITLEAKS_CONFIG")
  715. }
  716. if configPath != "" {
  717. if _, err := toml.DecodeFile(configPath, &config); err != nil {
  718. return fmt.Errorf("problem loading config: %v", err)
  719. }
  720. } else {
  721. _, err := toml.Decode(defaultConfig, &config)
  722. if err != nil {
  723. return fmt.Errorf("problem loading default config: %v", err)
  724. }
  725. }
  726. if len(config.Misc.Entropy) != 0 {
  727. err := entropyLimits(config.Misc.Entropy)
  728. if err != nil {
  729. return err
  730. }
  731. }
  732. if singleSearchRegex != nil {
  733. regexes["singleSearch"] = singleSearchRegex
  734. } else {
  735. for _, regex := range config.Regexes {
  736. regexes[regex.Description] = regexp.MustCompile(regex.Regex)
  737. }
  738. }
  739. whiteListCommits = make(map[string]bool)
  740. for _, commit := range config.Whitelist.Commits {
  741. whiteListCommits[commit] = true
  742. }
  743. for _, regex := range config.Whitelist.Files {
  744. whiteListFiles = append(whiteListFiles, regexp.MustCompile(regex))
  745. }
  746. for _, regex := range config.Whitelist.Regexes {
  747. whiteListRegexes = append(whiteListRegexes, regexp.MustCompile(regex))
  748. }
  749. for _, regex := range config.Whitelist.Repos {
  750. whiteListRepos = append(whiteListRepos, regexp.MustCompile(regex))
  751. }
  752. return nil
  753. }
  754. // entropyLimits hydrates entropyRanges which allows for fine tuning entropy checking
  755. func entropyLimits(entropyLimitStr []string) error {
  756. for _, span := range entropyLimitStr {
  757. split := strings.Split(span, "-")
  758. v1, err := strconv.ParseFloat(split[0], 64)
  759. if err != nil {
  760. return err
  761. }
  762. v2, err := strconv.ParseFloat(split[1], 64)
  763. if err != nil {
  764. return err
  765. }
  766. if v1 > v2 {
  767. return fmt.Errorf("entropy range must be ascending")
  768. }
  769. r := entropyRange{
  770. v1: v1,
  771. v2: v2,
  772. }
  773. if r.v1 > 8.0 || r.v1 < 0.0 || r.v2 > 8.0 || r.v2 < 0.0 {
  774. return fmt.Errorf("invalid entropy ranges, must be within 0.0-8.0")
  775. }
  776. entropyRanges = append(entropyRanges, r)
  777. }
  778. return nil
  779. }
  780. // getSSHAuth return an ssh auth use by go-git to clone repos behind authentication.
  781. // If --ssh-key is set then it will attempt to load the key from that path. If not,
  782. // gitleaks will use the default $HOME/.ssh/id_rsa key
  783. func getSSHAuth() (*ssh.PublicKeys, error) {
  784. var (
  785. sshKeyPath string
  786. )
  787. if opts.SSHKey != "" {
  788. sshKeyPath = opts.SSHKey
  789. } else {
  790. // try grabbing default
  791. c, err := user.Current()
  792. if err != nil {
  793. return nil, nil
  794. }
  795. sshKeyPath = fmt.Sprintf("%s/.ssh/id_rsa", c.HomeDir)
  796. }
  797. sshAuth, err := ssh.NewPublicKeysFromFile("git", sshKeyPath, "")
  798. if err != nil {
  799. if strings.HasPrefix(opts.Repo, "git") {
  800. // if you are attempting to clone a git repo via ssh and supply a bad ssh key,
  801. // the clone will fail.
  802. return nil, fmt.Errorf("unable to generate ssh key: %v", err)
  803. }
  804. }
  805. return sshAuth, nil
  806. }
  807. func (leak Leak) log() {
  808. b, _ := json.MarshalIndent(leak, "", " ")
  809. fmt.Println(string(b))
  810. }