nx_cleanview.nx source

1// nx_cleanview.nx -- one-shot CLI face of the ethical clean-serve engine, exposed as an MCP tool (nx_cleanview). 2// TWO modes: 3// nx_cleanview <url> -> the machine-readable SAFETY RECEIPT JSON (attacks neutralized + safe ads kept). 4// nx_cleanview resolve <url> -> F108d: sovereign GUARDED JS-EXEC media resolution. Fetches the page over the 5// SAME _best (Chrome-JA3) path clean-serve uses, then runs the page's inline JS 6// through the guarded sniff (fork + CPU/mem cap, NO network in the child) to 7// surface the runtime-constructed stream URL (VSK_HLS/DIRECT). This is the 8// _best-HTML-passthrough integration design (do NOT re-fetch with a weaker path). 9// External-<script src> _best-fetch enrichment = the next rung. 10// Shares the daemon's exact clean core (cs_clean_page / cs_receipt_json) so the receipt can NEVER drift from /receipt. 11// SSRF-guarded by construction (public http(s) only). Read-only, no persistent state, safe to fork-exec per call. 12// license_tier: ORIGINAL 13import "nx_syscalls.nx" 14import "nx_url_safety_lib.nx" // the ONE SSRF url ruler (us_url_safe / us_reason) 15import "nx_x509_trust_store.nx" 16import "nx_trust_store_load_from_certdata.nx" 17import "nx_https_fetch_follow.nx" 18import "nx_web_filter.nx" // wf_new / wf_seed (the block/allow data table) 19import "nx_clean_serve.nx" // cs_rc_new / cs_clean_page / cs_receipt_json / CR_* 20import "nx_video_sniff.nx" // vs_sniff_page_bundled (GUARDED) / vs_result_json / vs_extract_script_srcs / VSK_* 21import "nx_video_sniff_fetch.nx" // nx_hls_resolve_uri (relative-URL resolver) for external-bundle _best-fetch 22const NCV_MAGIC_8388608: i64 = 8388608 23const NCV_MAGIC_8192: i64 = 8192 24const NCV_MAGIC_8191: i64 = 8191 25const NCV_MAGIC_2560: i64 = 2560 26const NCV_MAGIC_65536: i64 = 65536 27const NCV_MAGIC_4194304: i64 = 4194304 28const NCV_MAGIC_1048576: i64 = 1048576 29const NCV_MAGIC_4000: i64 = 4000 30 31// -- minimal local helpers (mirror the daemon's csd_* / the core's cs_* idioms; kept local = self-contained) -- 32func ncv_put(dst: *u8, off: i64, s: *u8) -> i64 { var i: i64=0; while s[i]!=(0 as u8){dst[off]=s[i];off=off+1;i=i+1} return off } 33func ncv_putn(dst: *u8, off: i64, v: i64) -> i64 { var m: i64=v; if m==0 { dst[off]=48 as u8; return off+1 } let t: *u8=sys_mmap(24); var k: i64=0; while m>0 { t[k]=(48+(m%10)) as u8; m=m/10; k=k+1 } var j: i64=0; while j<k { dst[off]=t[k-1-j]; off=off+1; j=j+1 } return off } 34func ncv_slen(s: *u8) -> i64 { var n: i64=0; while s[n]!=(0 as u8){n=n+1} return n } 35func ncv_streq(a: *u8, b: *u8) -> i64 { var i: i64=0; while a[i]!=(0 as u8){ if a[i]!=b[i] { return 0 } i=i+1 } if b[i]!=(0 as u8) { return 0 } return 1 } 36func ncv_lc2(c: i64) -> i64 { if c>=65 { if c<=90 { return c+32 } } return c } 37func ncv_pfx_ci(s: *u8, sl: i64, p: *u8) -> i64 { var i: i64=0; while p[i]!=(0 as u8) { if i>=sl { return 0 } if ncv_lc2(s[i]&0xff)!=ncv_lc2(p[i]&0xff) { return 0 } i=i+1 } return 1 } 38func ncv_write_all(fd: i64, buf: *u8, n: i64) -> i64 { var off: i64=0; while off<n { let w: i64=sys_write(fd, ((buf as i64)+off) as *u8, n-off); if w<=0 { return off } off=off+w } return off } 39// body offset of a fetch result. DELEGATES to the ONE entity-body ruler in nx_clean_serve (cs_entity_body_off) -- the 40// same second-header-strip defect the daemon carried (a CRLF page lost everything before its first blank line, measured 41// 2026-09-02 on camwhores.tv); fixed in one place for both surfaces. 42func ncv_body_off(buf: *u8, n: i64) -> i64 { return cs_entity_body_off(buf, n) } 43// SSRF guard: a public fetch proxy MUST refuse private/loopback/internal (mirrors csd_url_safe exactly) 44// MIGRATED 2026-09-01 to the ONE SSRF ruler (see the same note in nx_clean_serve_daemon.nx). This body was a 45// byte-for-byte twin of csd_url_safe; the rules now live once, in nx_url_safety_lib, which also names WHICH 46// rule refused. Signature and return contract unchanged (1 = safe, 0 = refuse). 47func ncv_url_safe(url: *u8, ulen: i64) -> i64 { 48 return us_url_safe(url, ulen) 49} 50// F108d: guarded JS-exec media resolve on the _best-fetched HTML + external <script src> bundles (all fetched 51// over clean-serve's strong _best path -- the design proven this session; solves the standalone sniff's weak fetch). 52const NCV_SNIFF_CPU_SECS: i64 = 5 53const NCV_SRCCAP: i64 = 65536 // buffer for the extracted <script src> list 54const NCV_EXTCAP: i64 = 2097152 // concatenated external-bundle JS (2 MiB) 55const NCV_BUNCAP: i64 = 1048576 // one bundle response (1 MiB) 56const NCV_MAXSRC: i64 = 8 // fetch at most 8 external bundles (bounded; declared) 57func ncv_resolve(out: *u8, url: *u8, ulen: i64, store: *TrustStore) -> i64 { 58 let cap: i64 = NCV_MAGIC_8388608 59 let raw: *u8 = sys_mmap(cap) 60 let status: *i64 = sys_mmap(8) as *i64 61 // fetch over the SAME _best (Chrome-JA3) path clean-serve uses -- solves the weaker-fetch failure of the standalone sniff 62 let hn: i64 = nx_https_fetch_follow_best(url, store, raw, cap, 6, status) 63 if hn <= 0 { var o: i64 = ncv_put(out, 0, "{\"url\":\"" as *u8); o = ncv_put(out, o, url); o = ncv_put(out, o, "\",\"route\":\"fetch-failed\",\"status\":" as *u8); o = ncv_putn(out, o, status[0]); o = ncv_put(out, o, "}\n" as *u8); ncv_write_all(1, out, o); return 0 } 64 let boff: i64 = ncv_body_off(raw, hn) 65 let body: *u8 = ((raw as i64)+boff) as *u8 66 let blen: i64 = hn - boff 67 // EXTERNAL-BUNDLE ENRICHMENT: extract the page's <script src>, fetch each over _best, concat into extjs, so a 68 // framework player living in an external bundle can be run by the guarded sniff (not just inline scripts). 69 let srcs: *u8 = sys_mmap(NCV_SRCCAP) 70 let nsrc: i64 = vs_extract_script_srcs(body, blen, srcs, NCV_SRCCAP) 71 let extjs: *u8 = sys_mmap(NCV_EXTCAP) 72 let bundle: *u8 = sys_mmap(NCV_BUNCAP) 73 let absurl: *u8 = sys_mmap(NCV_MAGIC_8192) 74 let line: *u8 = sys_mmap(NCV_MAGIC_8192) 75 let bstatus: *i64 = sys_mmap(8) as *i64 76 var extlen: i64 = 0 77 var fetched: i64 = 0 78 var li: i64 = 0 79 var walking: i64 = 1 80 while walking == 1 { 81 if fetched >= NCV_MAXSRC { walking = 0 } else { 82 if (srcs[li] & 0xff) == 0 { walking = 0 } else { 83 var q: i64 = 0 84 var inl: i64 = 1 85 while inl == 1 { 86 let c: i64 = srcs[li] & 0xff 87 if c == 0 { inl = 0 } else { if c == 10 { inl = 0; li = li + 1 } else { if q < NCV_MAGIC_8191 { line[q] = c as u8; q = q + 1 } li = li + 1 } } 88 } 89 line[q] = 0 as u8 90 if q > 0 { 91 let rn: i64 = nx_hls_resolve_uri(url, ulen, line, q, absurl, NCV_MAGIC_8192) 92 if rn > 0 { 93 absurl[rn] = 0 as u8 94 let bn: i64 = nx_https_fetch_follow_best(absurl, store, bundle, NCV_BUNCAP, 6, bstatus) 95 if bn > 0 { 96 let bbo: i64 = ncv_body_off(bundle, bn) // skip the bundle's HTTP headers -> JS body only 97 var bp: i64 = bbo 98 while bp < bn { if extlen < (NCV_EXTCAP - 3) { extjs[extlen] = bundle[bp]; extlen = extlen + 1 } bp = bp + 1 } 99 if extlen < (NCV_EXTCAP - 3) { extjs[extlen] = 59 as u8; extlen = extlen + 1; extjs[extlen] = 10 as u8; extlen = extlen + 1 } // ';\n' separator 100 fetched = fetched + 1 101 } 102 } 103 } 104 } 105 } 106 } 107 extjs[extlen] = 0 as u8 108 // GUARDED JS-exec (fork + CPU/mem cap, no network in the child) on the page HTML + the _best-fetched external bundles. 109 let plan: *u8 = sys_mmap(NCV_MAGIC_2560) 110 let kind: i64 = vs_sniff_page_bundled(body, blen, extjs, extlen, plan, NCV_MAGIC_2560, NCV_SNIFF_CPU_SECS) 111 let jb: *u8 = sys_mmap(NCV_MAGIC_2560) 112 vs_result_json(kind, plan, jb, NCV_MAGIC_2560) 113 var o: i64 = ncv_put(out, 0, "{\"url\":\"" as *u8); o = ncv_put(out, o, url) 114 o = ncv_put(out, o, "\",\"js_media_resolve\":" as *u8); o = ncv_put(out, o, jb) 115 o = ncv_put(out, o, ",\"engine\":\"sovereign guarded JS-VM (fork+CPU/mem-cap, no-network child)\",\"scope\":\"page + up to " as *u8); o = ncv_putn(out, o, NCV_MAXSRC) 116 o = ncv_put(out, o, " external <script src> bundles, all fetched over _best (Chrome-JA3); bundles_fetched=" as *u8); o = ncv_putn(out, o, fetched) 117 o = ncv_put(out, o, "\"}\n" as *u8) 118 ncv_write_all(1, out, o) 119 return 0 120} 121func main(argc: i64, argv: *i64) -> i64 { 122 let out: *u8 = sys_mmap(NCV_MAGIC_65536) 123 if argc < 2 { let o: i64 = ncv_put(out, 0, "{\"error\":\"usage: nx_cleanview <url> | nx_cleanview resolve <url>\"}\n" as *u8); ncv_write_all(1, out, o); return 2 } 124 var mode_resolve: i64 = 0 125 var url: *u8 = argv[1] as *u8 126 if ncv_streq(argv[1] as *u8, "resolve" as *u8) == 1 { 127 if argc < 3 { let o: i64 = ncv_put(out, 0, "{\"error\":\"usage: nx_cleanview resolve <url>\"}\n" as *u8); ncv_write_all(1, out, o); return 2 } 128 mode_resolve = 1 129 url = argv[2] as *u8 130 } 131 let ulen: i64 = ncv_slen(url) 132 if ncv_url_safe(url, ulen) == 0 { let o: i64 = ncv_put(out, 0, "{\"safe\":false,\"error\":\"blocked-ssrf-guard\"}\n" as *u8); ncv_write_all(1, out, o); return 0 } 133 var r: i64 = nx_trust_store_load_from_certdata("data/mozilla_certdata.txt" as *u8, 512, NCV_MAGIC_4194304) 134 if r <= 0 { r = nx_trust_store_load_from_certdata("/volume1/homes/elderwesto/nishihost/data/mozilla_certdata.txt" as *u8, 512, NCV_MAGIC_4194304) } 135 if r <= 0 { let o: i64 = ncv_put(out, 0, "{\"safe\":false,\"error\":\"trust-store-unavailable\"}\n" as *u8); ncv_write_all(1, out, o); return 0 } 136 let store: *TrustStore = r as *TrustStore 137 if mode_resolve == 1 { ncv_resolve(out, url, ulen, store); return 0 } 138 // -- receipt mode (UNCHANGED: byte-identical behavior for the live MCP registration) -- 139 let t: *i64 = wf_new(); wf_seed(t) 140 let cap: i64 = NCV_MAGIC_8388608 141 let raw: *u8 = sys_mmap(cap) 142 let status: *i64 = sys_mmap(8) as *i64 143 let hn: i64 = nx_https_fetch_follow_best(url, store, raw, cap, 6, status) 144 if hn <= 0 { var o: i64 = ncv_put(out, 0, "{\"safe\":false,\"error\":\"fetch-failed\",\"status\":" as *u8); o = ncv_putn(out, o, status[0]); o = ncv_put(out, o, "}\n" as *u8); ncv_write_all(1, out, o); return 0 } 145 let boff: i64 = ncv_body_off(raw, hn) 146 let body: *u8 = ((raw as i64)+boff) as *u8 147 let blen: i64 = hn - boff 148 let rc: *i64 = cs_rc_new() 149 let sink: *u8 = sys_mmap(blen + NCV_MAGIC_1048576) 150 cs_clean_page(body, blen, t, sink, blen + NCV_MAGIC_1048576, rc) 151 let o: i64 = cs_receipt_json(rc, out, NCV_MAGIC_4000) 152 out[o] = 10 as u8 153 ncv_write_all(1, out, o+1) 154 return 0 155}