Skip to content

reconFTW is a tool designed to perform automated recon on a target domain by running the best set of tools to perform scanning and finding out vulnerabilities

26.8 KB
Raw Blame

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

[Unreleased]

Added

  • swagger_check() function in modules/web.sh: Swagger/OpenAPI definition discovery and authentication analysis using sj (Swagger Jacker). Three phases: (1) active brute-force discovery of hidden swagger/openapi definitions via sj brute with interlace parallelization, (2) aggregation of swagger URLs from nuclei (swagger-api/openapi templates) and SwaggerSpy OSINT output (with scope filtering via check_inscope), (3) endpoint extraction (sj endpoints) and authentication testing (sj automate) to identify unauthenticated API endpoints. Config: SJ_CHECK=true, SJ_BRUTE=true, SJ_AUTOMATE=true, SJ_TIMEOUT=30, SJ_MAX_TARGETS=200. Output: webs/swagger_urls.txt, vulns/swagger/endpoints_all.txt, vulns/swagger/automate/*.json, vulns/swagger/accessible_endpoints.txt.

Changed

  • Replaced metagoofil + exiftool with exifray in metadata() function (modules/osint.sh). Single tool handles both file discovery and metadata extraction, outputting structured JSON (osint/metadata.json) plus a human-readable summary (osint/metadata_results.txt).

Removed

  • Removed metagoofil from install.sh git repos, config/tool_versions.txt, and modules/core.sh tool checks.

Fixed

  • sanitize_domain() (modules/utils.sh) now URL-aware. Previously tr -cd 'a-zA-Z0-9.-' stripped :// from inputs like https://example.com, producing a garbage hostname (httpsexample.com) that poisoned the DNS/probe pipeline. The function now normalizes the input first (strip scheme, userinfo, path/query/fragment, port) and validates IPv4 octets inline (rejects https://1.2.3.999). The whitelist tr step is preserved as last-line injection defense. New bats cases cover scheme, path, port, userinfo, query/fragment, @ in path, and IPv4 URLs (valid + invalid).
  • Web mode (-w) accepts both -d <domain|url> and -l <file>. Previously -d ... -w errored out with "Web mode needs a website list file as target" because the -w dispatch branch only supported -l. The rewritten branch in reconftw.sh unifies both paths and errors only when neither is provided.
  • -l <file> -w pipeline now runs end-to-end. Before this fix, the -w branch copied the raw list to webs/webs.txt but webprobe_full reads from subdomains/subdomains.txt — which stayed empty — so probing skipped with SKIP_NOINPUT and cascaded skips through urlchecks, sub_js_extract, url_gf, url_ext. The branch now populates subdomains/subdomains.txt with sanitized hostnames (via _sanitize_list_entry), and truncates webs/webs.txt after start() so webprobe_full is the sole source of live URLs — downstream modules no longer mix raw user input (URLs with arbitrary paths/ports/dead hosts) with probed results.
  • webprobe_full and its helpers (_extract_probe_urls_from_json, _write_plain_webinfo) no longer discard all probe results in list mode. The jq scope filter select((.url // "") | contains($dom)) in modules/web.sh:213,225 matched the $domain value literally, which is "Multi" in -l runs — so every real URL was filtered out. Now bypassed when $dom == "Multi". Same fix applied to filter_in_scope_urls (lib/validation.sh) and the awk index($0, dom) gate in _write_plain_webinfo. webs/webs.txt, webs/webs_uncommon_ports.txt, webs/web_full_info.txt, and webs/web_full_info_plain.txt are now properly populated in list mode.
  • -w mode no longer leaks stale state across runs, nor clobbers enumeration data from prior non-web runs. The shared output dir (Recon/Multi for -l, Recon/<domain> for -d) is reused between invocations, and many webs_menu modules persist via anew / >> across a wide set of paths (webs/webs_all.txt, webs/url_extract*, webs/virtualhosts.txt, .tmp/* caches, root-level assets.jsonl, js/, fuzzing/, gf/, screenshots/, nuclei_output/, cms/, vulns/, subdomains/subtakeover.txt, subdomains/* from sub_js_extract/tls_ip_pivots, hosts/grpc_reflection.txt) or skip on checkpoint (.called_fn/.<func>). A new helper prepare_web_mode_scope() in modules/modes.sh differentiates by input mode to match each scope's semantics:
    • -l Multi mode: the shared dir is only ever populated by -l -w invocations, so nothing in it predates a web run. The helper makes a read-only snapshot of the input list via cp "$flist" "$_list_snapshot" (tmpfile outside $dir) up front, then wipes every file under $dir except cross-run tracking (.log/*, .incremental/*, .called_fn/*, debug.log) plus $flist itself (belt-and-suspenders for when the list path falls inside $dir, e.g. -l Recon/Multi/scope.txt). All subsequent operations — including CRLF stripping via bash parameter expansion ${_t%$'\r'} — run against the snapshot. The user's input file is provably untouched post-snapshot: content md5, mtime, and inode all survive a run. This catches current and future output dirs (nuclei_output/, cms/, virtualhosts/, any new module's outputs) without requiring an explicit allowlist to stay in sync. subdomains/subdomains.txt is written with sort -u > from the sanitized snapshot — each list defines its own scope.
    • -d <domain> mode: the per-domain dir may hold prior -a/-s/-p recon state that must survive (subdomain enumeration, hosts/ips.txt and hosts/origin_ips.txt needed by virtualhosts()/tls_ip_pivots(), nuclei_output/** history, osint/*, etc.). The helper wipes only pure web-pipeline artifacts (webs/*, .tmp/*, js/*, fuzzing/*, gf/*, screenshots/*, assets.jsonl) and preserves the rest. The target is appended to subdomains/subdomains.txt via anew -q so enumerated subs stay in scope. The one-shot -w dispatch also sets DIFF=true, matching the global DIFF=true that monitor_mode already sets. Both the one-shot -w dispatch and every monitor_mode -w -l cycle now call the same helper and check its return status — if the new scope can't be validated (mktemp/cp failed, or every list entry was rejected by sanitization so the staged hostname list ended up empty), the helper aborts before wiping $dir so prior state is preserved, the one-shot path exits with a clear error, and the monitor path skips webs_menu for that cycle only and falls through to the normal post-case flow (monitor_snapshot, MONITOR_MAX_CYCLES check, sleep $interval_sec). An earlier continue branch was removed because it bypassed those checks and busy-looped on persistent prep failures. Consecutive -w runs produce fully scoped output with zero cross-run contamination while preserving INCREMENTAL_MODE baselines and prior enumeration coverage.

[v4.1] - 2026-03-06

Added

Advanced Subdomain & Asset Discovery

  • sub_srv function: SRV record enumeration (_ldap._tcp, _sip._tcp, _xmpp-server._tcp, etc.) via dnsx. Discovers service infrastructure hosts. Config: SRV_ENUM=true. Output: subdomains/srv_records.txt. Data file: data/wordlists/srv_prefixes.txt (~27 SRV prefixes).
  • sub_ptr_cidrs function: PTR sweep over ASN CIDR ranges discovered by sub_asn. Expands CIDRs via mapcidr, runs reverse PTR lookups via dnsx, filters in-scope results. Config: PTR_SWEEP=false (off by default), PTR_SWEEP_MAX_IPS=50000 (safety limit). Output: subdomains/ptr_pivots.txt.
  • sub_ns_delegation function: discovers delegated DNS zones (subdomains with their own NS records) and attempts AXFR zone transfers on each delegated nameserver. Complements existing zonetransfer() which only checks the main domain's NS. Config: NS_DELEGATION=true. Output: subdomains/ns_delegated_zones.txt.
  • tls_ip_pivots function: TLS certificate harvesting from raw IPs (not just known subdomains like sub_tls). Three phases: (A) passive cert harvest extracting SAN/CN via tlsx -json, (B) SNI brute-force using discovered labels as candidates (DEEP mode only), (C) reverse PTR + SNI probing via tlsx -rev-ptr-sni. Includes delta-only httpx probe for newly discovered subdomains. Config: TLS_IP_PIVOTS=false, TLS_IP_SNI_BATCH_SIZE=1000, TLS_IP_DELTA_PROBE=true. Output: hosts/tls_ip_certs.jsonl, subdomains/tls_ip_pivots.txt.
  • sub_js_extract function: extracts hostnames from JS/crawl output files (js/nojs_links.txt, js/js_livelinks.txt, webs/url_extract.txt, js/js_secrets.txt, etc.), resolves new candidates, and delta-probes them. Creates a feedback loop from web analysis back into subdomain discovery. Config: JS_SUB_EXTRACT=true.
  • well_known_pivots function: checks well-known metadata endpoints (/.well-known/security.txt, /.well-known/openid-configuration, /.well-known/oauth-authorization-server) for hostname references matching the target domain. Config: WELLKNOWN_PIVOTS=false, WELLKNOWN_MAX_TARGETS=200.
  • Reactivated virtualhosts() in all scan modes (recon, subs_menu, webs_menu, zen_menu, multi-domain recon). Was previously commented out. Gated by existing VIRTUALHOSTS=false config flag.

Pipeline Integration

  • Subdomain Phase 4 (parallel): added sub_srv and sub_ptr_cidrs alongside sub_noerror and sub_dns.
  • Subdomain Phase 5 (parallel): added sub_ns_delegation alongside sub_tls and sub_analytics.
  • Web Detection: added tls_ip_pivots and virtualhosts after portscan/geo_info in both parallel and sequential paths.
  • Web Analysis: added sub_js_extract and well_known_pivots after jschecks.
  • passive() mode: PTR_SWEEP, NS_DELEGATION, and TLS_IP_PIVOTS are saved/restored as false to prevent active techniques in passive-only mode.

Config Profiles

  • Full profile (reconftw_full.cfg): enables PTR_SWEEP, SRV_ENUM, NS_DELEGATION, TLS_IP_PIVOTS, JS_SUB_EXTRACT, WELLKNOWN_PIVOTS.
  • Quick profile (reconftw_quick.cfg): enables only SRV_ENUM (lightweight); disables all others.
  • Stealth profile (reconftw_stealth.cfg): enables only SRV_ENUM; disables all active techniques.

[v4.0] - 2026-02-20

Added

  • nuclei_dast module: dedicated nuclei -dast pass over webs/, url_extract, and gf/ candidates (outputs nuclei_output/dast_json.txt, vulns/nuclei_dast.txt).
  • Portscan strategy PORTSCAN_STRATEGY=naabu_nmap: naabu pre-discovery + targeted nmap outputs (hosts/naabu_open.txt, hosts/portscan_active_targeted.*).
  • Optional UDP scan: PORTSCAN_UDP=true produces hosts/portscan_active_udp.*.
  • Monitor alert threshold MONITOR_MIN_SEVERITY (critical|high|medium|low|info).
  • SSRF alternate protocol payload pass via config/ssrf_payloads.txt (output vulns/ssrf_alt_protocols.txt).
  • Katana headless profile selection: KATANA_HEADLESS_PROFILE=off|smart|full.
  • DEEP-only ffuf recursion: FUZZ_RECURSION_DEPTH.
  • CDN origin discovery (best-effort) via hakoriginfinder with CDN_BYPASS=true (output hosts/origin_ips.txt).
  • Password dictionary engine cewler (Python) with tuning knobs.
  • Adaptive permutation wordlist: PERMUTATIONS_WORDLIST_MODE=auto|full|short with PERMUTATIONS_SHORT_THRESHOLD=100. Uses a curated 162-word list for large targets (>100 subs) and the full 849-word list for small targets or DEEP mode.
  • Auto-detection of DNS resolver: DNS_RESOLVER=auto|puredns|dnsx. Two-step heuristic: (1) public local IP → puredns, (2) cloud metadata endpoint (169.254.169.254) → puredns, otherwise → dnsx. Handles cloud VPS with private IPs (AWS, GCP, DO, etc.) correctly. Configurable with DNSX_THREADS and DNSX_RATE_LIMIT.
  • Global DNS resolution wrappers _resolve_domains() and _bruteforce_domains() in modules/utils.sh — all 16 puredns call sites now use these wrappers for consistent resolver selection.
  • Resolver hardening for subdomain brute/resolve: fail-fast when required resolver files are missing, stricter resolver downloads (RESOLVER_DOWNLOAD_*), and optional DNS hard-timeouts (DNS_BRUTE_TIMEOUT, DNS_RESOLVE_TIMEOUT, where 0 disables timeout) with heartbeat progress (DNS_HEARTBEAT_INTERVAL_SECONDS).
  • Cross-platform local IP detection via _get_local_ip() using default route (works on any macOS/Linux interface naming).
  • DNS permutation benchmark script benchmark_dns_permutations.sh for comparing gotator, alterx, and ripgen.
  • Built-in wordlists and patterns are now vendored in-repo under data/ (wordlists: data/wordlists/, patterns: data/patterns/). Config exposes DATA_DIR, WORDLISTS_DIR, and PATTERNS_DIR.
  • --no-banner to disable banner output (banner is on by default).
  • --no-report to skip report generation/exports.
  • --gen-resolvers to generate resolvers with dnsvalidator.
  • --force to ignore cached module markers and re-run all phases.

Modular Architecture

  • Split monolithic script into focused modules under modules/:
    • core.sh — lifecycle, logging, notifications, cleanup
    • modes.sh — scan modes and argument parsing
    • subdomains.sh — subdomain enumeration
    • web.sh — web analysis, fuzzing, nuclei scans
    • vulns.sh — vulnerability scanning
    • osint.sh — OSINT functions
    • axiom.sh — Axiom fleet helpers
    • utils.sh — shared utilities
  • Validation library at lib/validation.sh for input sanitization.
  • Shared libraries in lib/common.sh and lib/parallel.sh.

New CLI and Runtime Modes

  • --health-check for system diagnostics.
  • --dry-run to preview commands.
  • --parallel to run independent functions in parallel.
  • --no-parallel to force sequential mode.
  • --source-only for sourcing/testing workflows.
  • Continuous monitoring mode:
    • --monitor
    • --monitor-interval <minutes>
    • --monitor-cycles <n> (0 = infinite)
  • --report-only to rebuild report artifacts from existing outputs.
  • --refresh-cache to force cache refresh.
  • --export json|html|csv|all export pipeline.

Reporting and Output

  • Consolidated reporting artifacts:
    • report/report.json
    • report/index.html
    • report/latest/report.json
    • report/latest/index.html
  • Export artifacts:
    • JSONL: report/findings_normalized.jsonl, report/export_all.jsonl
    • CSV: report/subdomains.csv, report/webs.csv, report/hosts.csv, report/findings.csv
  • Delta sections and timeline in report outputs.
  • Monitor history snapshots under .incremental/history/<timestamp>/.
  • Per-cycle alert summary at .incremental/history/<timestamp>/alerts.json.
  • Machine-readable run summary at .log/perf_summary.json.
  • Structured JSON logging support (STRUCTURED_LOGGING=true) with function/subfunction start/end events.

AI and Automation Enhancements

  • AI report generation integrated with local models via reconftw_ai.
  • AI profiles: executive, brief, bughunter.
  • Structured AI output artifact: ai_result/reconftw_analysis.json.
  • Human AI report artifact: ai_result/reconftw_analysis_<profile>_<timestamp>.md|txt.
  • New AI configuration controls:
    • AI_EXECUTABLE
    • AI_PROMPTS_FILE
    • AI_MAX_CHARS_PER_FILE
    • AI_MAX_FILES_PER_CATEGORY
    • AI_REDACT
    • AI_ALLOW_MODEL_PULL
    • AI_STRICT

Configuration and Profiles

  • New profile configs:
    • config/reconftw_quick.cfg
    • config/reconftw_full.cfg
    • config/reconftw_stealth.cfg
  • Port lists externalized for easier maintenance:
    • config/tls_ports.txt
    • config/uncommon_ports_web.txt
  • Typed cache TTL controls:
    • CACHE_MAX_AGE_DAYS_RESOLVERS
    • CACHE_MAX_AGE_DAYS_WORDLISTS
    • CACHE_MAX_AGE_DAYS_TOOLS

Recon and Enumeration Enhancements

  • ASN enumeration via sub_asn() using asnmap.
  • New ASN_ENUM option and outputs:
    • hosts/asn_numbers.txt
    • hosts/asn_cidrs.txt
  • Deep wildcard detection (DEEP_WILDCARD_FILTER) based on iterative wildcard checks.
  • Time-based certificate filtering (DNS_TIME_FENCE_DAYS).
  • Sensitive domain exclusion (EXCLUDE_SENSITIVE, config/sensitive_domains.txt).
  • New integrated engines/tools:
    • toxicache for additional web cache poisoning coverage.
    • postleaksNg in apileaks for Postman leak discovery.
    • favirecon as favirecon_tech for favicon-based technology reconnaissance.
    • Praetorian integrations:
      • fingerprintx for service fingerprinting after host port discovery (hosts/fingerprintx.[jsonl|txt]).
      • brutus as optional spraying engine (SPRAY_ENGINE=brutus) with DEEP gating support.
      • titus and noseyparker as selectable GitHub repo secret engines (SECRETS_ENGINE).
      • gato as optional GitHub Actions audit module (github_actions_audit, output osint/github_actions_audit.[json|txt]).
      • julius as optional LLM endpoint probing module (llm_probe, output webs/llm_probe.[jsonl|txt]).
    • TInjA as the default ssti engine (SSTI_ENGINE=TInjA, with legacy fallback).
  • second-order as default brokenLinks engine (BROKENLINKS_ENGINE=second-order, with legacy fallback).
  • ghleaks in github_leaks for GitHub-wide secret search across all public repositories (combines GitHub Code Search API with gitleaks detection rules). Configurable via GITHUB_LEAKS, GHLEAKS_THREADS, and --exhaustive mode in DEEP.

Changed

  • PERMUTATIONS_ENGINE is now treated as deprecated/ignored legacy config; gotator is always used for permutations.
  • Removed obsolete/no-op knobs from public config examples and templates: NUCLEI_FLAGS, NUCLEI_FLAGS_JS, BRUTESPRAY_THREADS, METAFINDER_LIMIT, RESOLVE_DOMAINS_THREADS, INTRUSIVE, FARADAY_SERVER, FARADAY_USER, FARADAY_PASS.
  • Removed no-op legacy overrides from lightweight profile examples (BROKENLINKS_ENGINE/SSTI_ENGINE when module is disabled).

Progress and UX

  • Progress/ETA functions (progress_init(), progress_step()) integrated in recon flows.
  • Better quick-rescan and monitor UX with adjusted totals and explicit cycle behavior.

Testing and Tooling

  • Expanded BATS test coverage (unit/integration/security) and mocks.
  • Integration tests for full flow and checkpoint behavior.
  • Additional test helpers/scripts for artifact and mode validation.

CI/CD and Docker

  • ShellCheck failures now fail CI (removed permissive continue-on-error).
  • Added Docker nightly verification steps (--help, --health-check).
  • Nightly image health-check failures now fail workflows.
  • Docker image security improved by generating Axiom SSH keys at runtime via Docker/entrypoint.sh instead of build-time.

Changed

  • Fixed webprobe_simple() httpx merge bug (probe output now correctly merged with prior cache).
  • Wrapped local wafw00f execution with run_command for dry-run/log consistency.
  • PORTSCAN_ACTIVE_OPTIONS default no longer includes --script vulners; moved to PORTSCAN_DEEP_OPTIONS.
  • ssti: if TInjA is missing, the module warns and skips (no automatic ffuf fallback).
  • iishortname: uses shortscan only (sns removed).
  • nuclei_dast: when vuln scanning is enabled (VULNS_GENERAL=true, e.g. -a), the DAST pass is force-enabled to avoid accidental coverage loss.
  • Permutation engine simplified to gotator-only (benchmark showed gotator produces the best quality permutations with exclusive valid findings that other tools miss).
  • DNS resolver auto-selection is evaluated once per run and cached (see DNS_RESOLVER_SELECTED via init_dns_resolver()).
  • Main script reduced to entry-point style with modular delegation.
  • Module loading order updated to include shared libraries (lib/common.sh, lib/parallel.sh).
  • Replaced repetitive utility patterns with centralized helpers:
    • ensure_dirs()
    • skip_notification()
    • count_lines()
    • run_tool()
    • should_run_function()
  • Profile configuration behavior changed to override-only semantics (preserving preloaded secrets/env).
  • Nuclei targeting adjusted to prioritize URL-with-protocol inputs and reduce duplicate host scans.
  • Parallel scheduler improved with backpressure signals from adaptive rate limiting.
  • Threading/performance tuning improved with PERF_PROFILE (low|balanced|max).
  • Output header now uses a banner-style H1 block with rules sized to the header text.
  • Default exports now generate all report artifacts unless --no-report is set.
  • Help output formatting uses proper color escapes (no literal \033).
  • --pretty option removed (no longer supported).
  • Health-check tool list updated to reflect actual runtime usage (nmap, nmapurls).
  • Fixed duplicate findings/reporting edge cases.
  • Corrected sub_asn() behavior to avoid redundant repeated runs.
  • sub_asn() now requires PDCP_API_KEY before running asnmap, logs explicit skip reason when unset, and applies a 120s timeout guard when available.
  • Corrected host matching in WAF/slow-host grouping to literal-safe comparisons.
  • Parallel group failures now propagate correctly instead of being silently ignored.

Cross-platform/Operational Fixes

  • Improved macOS compatibility for stat, df, and sed behavior.
  • Improved cache lifecycle behavior and runtime refresh handling.
  • Progress totals now correctly account for quick-rescan skips and execution model differences.

Fixed

Stability and Correctness

  • List mode (-l) now processes all targets reliably (FD-safe loops).
  • Fixed syntax/runtime edge cases in module scripts.
  • Corrected timing/perf summary handling edge cases.
  • Resolved readonly/re-source issues in script lifecycle.
  • .log/*.txt files are no longer empty (LOGFILE no longer overwritten by debug log).
  • Incidents no longer show \n literals; debug log line only appears if non-empty.
  • Passive mode no longer runs webprobe_simple/webprobe_full.
  • Missing webs/webs_new.txt now handled safely.

Dry-Run Mode

  • Wrapped 83 unwrapped command executions across all modules to respect DRY_RUN flag:
    • osint.sh: 24 commands (whois, dig, curl APIs, gitdorks_go, enumerepo, porch-pirate, postleaksNg, misconfig-mapper, exiftool, interlace git clone, Python tools: dorks_hunter, metagoofil, SwaggerSpy, EmailHarvester, LeakSearch, msftrecon, Scopify, Spoofy)
    • subdomains.sh: 15 commands (asnmap, dig zone transfer, curl APIs, hakip2host, csprecon, regulator, cloud enumeration)
    • web.sh: 20 commands (fav-up, favirecon, grpcurl, CMSeeK, JSA, getjswords, jsluice, wget, interlace/ffuf, curl WebSocket, pydictor, second-order)
    • vulns.sh: 6 commands (Corsy, Oralyzer, interlace/ffuf LFI, TInjA/legacy SSTI, interlace/ghauri SQLi, toxicache)
    • core.sh: 8 commands (tar/curl uploads, Telegram/Discord/Slack notifications, connectivity checks)
    • modes.sh: 8 commands (faraday-cli reporting)
    • axiom.sh: 3 commands (hakip2host, mapcidr)
  • Added module-level dry-run command tracking and summaries in lib/ui.sh:
    • ui_dryrun_track(): Records commands during dry-run execution
    • ui_dryrun_summary(): Displays count + unique tool names at module completion
    • ui_dryrun_reset(): Clears tracking at module boundaries
  • Dry-run summaries now show:
    • Normal mode: [DRY-RUN] Would execute N commands: tool1, tool2, ...
    • Verbose mode: Full command list with normalized single-line formatting
    • Quiet mode: No dry-run output
  • Module lifecycle hooks (_print_module_start(), _print_module_end()) now reset and display dry-run summaries automatically
  • Command normalization removes newlines/extra whitespace from multi-line command strings for cleaner display
  • Dry-run command previews now redact sensitive CLI values (for example -token-string, API keys, and bearer tokens).
  • Dry-run in parallel mode now aggregates commands emitted from child subshells, so module summaries reflect full command volume.
  • Dry-run no longer produces noisy parser/file errors in guarded paths (apileaks, sub_crt, prototype_pollution, fuzzparams, sub_tls, sub_permut, urlchecks, favirecon_tech, webcache, brokenLinks, ssti).
  • test_ssl now respects run_command, preventing real execution during dry-run.
  • sub_asn now applies timeout fencing with kill semantics (-k) when timeout tooling is available and reports non-OK outcomes correctly.
  • subdomains finalization now safely creates missing incremental files before counting to avoid stderr noise.
  • Fixed benign non-zero paths (anew, optional pipelines) that were surfacing as false ERR(1) noise in logs.
  • Hardened sub_crt JSON parsing to avoid noisy jq parse errors on malformed/empty upstream responses.
  • Parallel job output now strips ANSI/control sequences before replaying tail/full logs, reducing report/export parser issues.
  • Function/subfunction end log timestamps now use real end times (not reused start timestamps).
  • Added heartbeat progress lines for long-running phases (nuclei, ffuf/interlace, cmseek, and selected passive collectors) to avoid "stuck" perception.
  • Added run_with_heartbeat_shell for complex shell/redirection cases to keep heartbeat coverage in long-running URL collection paths.
  • Skip/cache output now includes explicit reason hints (config, noinput, cache) while preserving existing summary counters.
  • Parallel UI now defaults to a cleaner mode with aggregated progress snapshots (X/Y, %, elapsed, ETA) and reduced process-noise.
  • In clean parallel UI mode, Queue lines are suppressed when pending count is 0 to keep terminal output compact.
  • Added configurable parallel UX and sizing knobs in reconftw.cfg (PARALLEL_UI_MODE, progress visibility toggles, and per-group parallel sizes).
  • Parallel batch summaries now include completion ratio (completed/total) to make long blocks easier to track.

Parallel Execution and Status Reporting

  • Parallel aggregation now preserves per-function final badges (OK/WARN/FAIL/SKIP/CACHE) from module functions/subfunctions instead of inferring only from process exit code.
  • Parallel batch summary lines now expose real counters for ok/warn/fail/skip/cache.
  • Cache/skip messaging was normalized to avoid unformatted "already processed" lines and keep status output consistent.

Removed

  • Obsolete tool-based modules: cors (Corsy), open_redirect (Oralyzer), prototype_pollution (ppmap), favicon real-IP discovery (fav-up).
  • Deprecated/removed config flags: CORS, OPEN_REDIRECT, PROTO_POLLUTION, FAVICON.
  • Permutation engines ripgen, alterx, and PERMUTATIONS_ENGINE=both option. Gotator is now the only permutation engine (benchmark winner: best yield, exclusive findings, fastest).
  • Legacy PERMUTATIONS_OPTION alias.

Migration Notes

  • Prefer Nuclei templates (and nuclei_dast) for CORS/open-redirect/prototype-pollution coverage.
  • PERMUTATIONS_ENGINE now only accepts gotator. Remove any ripgen/alterx/both values from custom configs.
  • If running from home/behind NAT, the resolver auto-detection will use dnsx automatically. Set DNS_RESOLVER=puredns to force puredns on VPS.

Security

  • Removed unsafe eval usage on user-influenced input paths.
  • Hardened input sanitization against shell metacharacter injection.
  • Improved quoting and command construction across modules.
  • Secrets handling hardened (secrets.cfg checks, env fallback patterns, log redaction with command tracing).
  • Docker key handling moved to runtime generation to avoid shared-key risk across images.