Boyayyun iri (Seeds) da ke bayan kwanukan NIST ba Hashes ba ne tun fil azal, kuma Claude ya gano hakan a kasa da rabin sa'a

Tambayar da ta ci gaba da zama a bude bayan mutum daya tilo da zai iya amsa ta ya rasu a shekarar 2023 tana da amsa mai sauki: na'urar samar da lambobi masu bazuwa ta shekarun 1990 da aka fara daga agogo. Marubutan sun ce wannan amsar tana kara musu kwarin gwiwa cewa an zabi kwanukan cikin gaskiya, ba ragewa ba.

Takaitaccen Bayani na Zartarwa

• NIST ta karbi ECDSA a shekarar 2000 a matsayin wani bangare na FIPS 186-2 Digital Signature Standard. Bayanan ka'idar sun wallafa sigogin "iri" (seed) wadanda darajar hash dinsu aka nufa don kayyade lambar kwanar b (curve coefficient b). Hukumar NSA ce ta samar da irin ba tare da bayani ko hujja ba.
• A watan Oktoba na 2026, wani ma'aikacin Anthropic ya tambayi wani samfurin bincike na cikin gida don gano yadda aka samar da irin, ba tare da nuna alamar inda za a duba ba. A kasa da rabin sa'a, Claude ya gano cewa babu ko daya daga cikin irin da yake sakamakon hash. Sakamako ne na lrand48, wata na'urar samar da lambobi masu kama da bazuwa (pseudorandom number generator) da aka saba amfani da ita wadda aka fara daga agogo.
• Abin da aka gano ba karya tsarin sirri ba ne, ko sabon hari, ko rauni a cikin kwanukan ECDSA. Marubutan sun ce yana kara musu kwarin gwiwa cewa an zabi kwanukan ba tare da kofar baya (backdoors) ko rauni na gangan ba.
• Claude ya kuma bayyana maki na tushe (base points) na kwanuka shida daga Mayu 1997 a matsayin wadanda suka fito daga na'urar samar da lambobi masu bazuwa a cikin FreeLIP, kunshin lissafin manyan lambobi na Arjen Lenstra. A iya sanin marubutan, ba a taba bayyana wani maki na tushe na NIST ba kafin wannan.
• Tambayoyi masu bude sun rage: yadda aka zabi maki na tushe na P-224, P-256, P-384 da P-521, da kuma wadanne sharudda, idan akwai, ke bayan kwanukan da aka tsallake.

Lambar kwanar da ta yi kama da an samo ta daga wani abu da kuma irin da babu wanda zai iya bayyanawa #

NIST ta karbi ECDSA a shekarar 2000 a matsayin wani bangare na FIPS 186-2 Digital Signature Standard. Bayanan ka'idar don kwanukanta sun dauki sigar "iri" (seed) wadda ake amfani da darajar hash dinta don kayyade lambar b (coefficient b) ta kwanar. A kan takarda, wannan tsari ne mai kwantar da hankali. Idan lambar ta fito daga hash na irin da aka wallafa, kamar yadda hujjar ke cewa, babu wanda zai iya saita kwanar da hannu.

Wannan hujjar tana da inganci ne kawai gwargwadon ingancin irin. Hukumar NSA ce ta bayar da irin ba tare da bayani ko hujja ba, kuma wannan gibin shi ne daukacin matsalar. Hash kofa ce mai hanya daya: yana tabbatar da cewa lambar ta biyo bayan irin, amma ba ya cewa komai game da inda irin ya fito. Tsawon shekaru irin sun zauna a cikin ma'aunin a matsayin tsayayyun lambobi da ba a bayyana su ba.

Akwai labarin da aka kirkira don cike gibin. Sigar da ta fi shahara ta nuna cewa Jerry Solinas, wani ma'aikacin NSA da ya dade, ya samar da irin ne daga hash na wata kalmar sirri mai ban dariya. Tushen bayanin daga baya ya ce wannan ya tabbata ba gaskiya ba ne. Labarai irin wannan ana cika yarda da su domin suna da takamaiman bayani kuma suna da dan ban sha'awa, kuma saboda babu wanda zai iya karyata su. Suna gazawa a matsayin shaida. Sakamakon zahiri shi ne babu wanda zai iya gwada ikirarin, kuma binciken neman jimlar, wanda aka gina daga jita-jita, ya tashi a tutar babu.

Mutumin da ya kamata a tambaya a fili shi ne Solinas. Jerry ya rasu a shekarar 2023, kuma sauran abokan aiki a wancan lokacin ba su ba da amsa ba. Hakan ya bar tambayar yadda aka samo wadannan irin tun asali a bude.

Zan iya cewa wannan ita ce hanyar da ta dace a kalli yanayin: tambayar asali (provenance) da ta dogara da tunanin 'yan mutane kadan tsari ne mai rauni ga ma'aunin da ke tallafawa sa hannu na dijital. Abu mai ban sha'awa game da abin da ya faru a gaba shi ne bai bukaci wadannan tunanin ba. Amsar tana zaune a cikin irin da aka wallafa, tana jiran wani ya gwada zaton da ya dace.

Minti talatin, babu alama, da sakamakon da ba hash ba ne #

A watan Oktoba na 2026, wani ma'aikacin Anthropic mai son sani ya tambayi wani samfurin bincike na cikin gida ko zai iya gano yadda aka samar da wadannan irin, ba tare da nuna alamar inda za a duba ba. Yadda aka tsara kalaman yana da muhimmanci. Ba a gaya wa samfurin abin da zai nema ba, kuma ba a mika masa amsar don ya tabbatar ba.

A kasa da rabin sa'a, Claude ya gano asalin irin NIST P-192, P-224, P-256, P-384 da P-521. Sakamakon ya fito gatsau: babu ko daya daga cikin irin da yake sakamakon wata manhajar hash. Sakamako ne na na'urar samar da lambobi masu bazuwa ta lrand48 a dakin karatu na C.

Don haka labarin kalmar sirri ba za a taba gano shi ba, domin babu wata kalmar sirri da za a nema. An samar da irin ne ta hanyar na'urar samar da lambobi masu kama da bazuwa ta yau da kullun da aka fara daga agogo. Bayani ne marar kwalliya, kuma wannan shi ne ainihin abin sha'awarsa: ba ya bukatar asiri, ba ya bukatar ilimi na musamman, kuma babu tsayayyiyar lamba da ba a bayyana ba.

Tsarin yana da saukin bayyanawa. Kowace maimaitawa (iteration) ta algorithm tana ciro bytes 20 daga na'urar samarwa a matsayin irin da ake gwadawa (candidate seed). Adadin maimaitawar shi ne bayani mai ban sha'awa, domin yana nuna cewa aikin bincike ne ba cirowa daya tak ba. P-256, alal misali, ya bukaci maimaitawa 1842 don fitar da daidaitaccen irin. Duk wanda ya gudanar da wannan shirin ya bar shi ya yi ta jujjuya wadanda ake gwadawa har sai da daya ya samar da kwanar da ta cika sharuddansu.

Marubutan suna da taka-tsantsan game da abin da wannan ke nufi da wanda ba ya nufi, kuma yadda suka gabatar da abin ya cancanci a maimaita. Wannan ba karya tsarin sirri ba ne, ko sabon hari, ko rauni a cikin kwanukan ECDSA. Babu wani mataki da za a dauka. Sun kara mataki daya gaba suka ce yana kara musu kwarin gwiwa cewa an zabi kwanukan ba tare da wata kofar baya ko rauni na gangan ba. Wannan kammalawa ce mai dauke da sharadi, kuma ya kamata ta ci gaba da zama haka. Tushen bayanin ya gabatar da ita a matsayin kwarin gwiwar marubutan, wanda abu ne daban da hujjar rashin laifin kwanukan. Abin da gano hakan ke bayarwa shi ne bayani na yau da kullun, mai iya gwadawa a madadin labarin da ba za a iya tabbatarwa ba.

Yadda na'urar duba ta gano sawun yatsan na'urar samarwa ta shekarun 1990 #

Hanyar da aka bi ta cancanci kulawa, domin ita ce bangaren da wasu mutane za su iya sake amfani da shi.

Claude ya yi tsammanin cewa manhajar da ta samar da kwanukan ta yi amfani da PRNG wanda ya zo tare da dakunan karatu na C na shekarun 1990. Tushen bayanin yana daukar wannan a matsayin zato, kuma zato ne mai kyau. Ya takaita bincike marar iyaka zuwa ga gajeren jerin wadanda ake gwadawa daga zamani guda.

Daga nan sai Claude ya rubuta na'urar duba (scanner). Ta zame a kan kowane irin da aka wallafa bit daya bayan daya kuma ta gwada ko yankunan da ke kusa da juna sakamako ne na jere na na'urar samarwa ta zamanin shekarun 1990. Dubawar ta rufe hade-hade 34 na na'urar samarwa da yankan sakamako. Nasarar ta fito a matsayin tsari a cikin bits na sama, wanda zai iya faruwa da kansa bisa kuskure da yiwuwar 1/5810.

Wannan tsarin alama ce ta linear congruential generator mai suna lrand48, wanda ke dawo da bits 31 na sama kawai na yanayi mai bits 48 (48-bit state). A nan ne tsarin na'urar samarwar ke tona asiri. Saboda lrand48 yana boye bits 17 na yanayinsa mai bits 48 a kowane kira, sakamakon yana da tsarin daidaito wanda jeren lambobi na bazuwa na gaske ba zai nuna ba, kuma na'urar duba da aka gina don neman wannan daidaiton za ta iya gano shi.

Don dawo da yanayin, Claude ya dauki kalmar farko (first word) kuma ya gwada dukkan darajoji 2^17 na bits 17 da aka boye. Wannan karamin bincike ne, kuma yana da arha domin na'urar samarwar tana da sauki sosai. Kowane yanayi da ake gwadawa ana iya duba shi a kan yanki na gaba na irin, kuma wanda ya yi daidai ya sa daukacin jerin suka hau layi.

Irin kwanukan lambobin farko (prime-curve seeds) sun fi wahala, ko da yake kadan ne kawai. Sun fito ne daga wannan na'urar samarwa ta lrand48 amma sun rike bits kadan na kowane sakamako. Karancin bits da ake gani yana nufin karancin abin da za a dora zato a kai, don haka dabarar gwada duka (brute-force) ba ta aiki sosai a nan. Amsar tushen bayanin ita ce dawo da yanayin linear congruential generator da aka rage tsayinsa matsala ce ta lattice ta littafin karatu, wadda ake warwarewa da manhajar rage tushe ta Lenstra-Lenstra-Lovasz (LLL). Samfuri zai iya gane hakan, kuma ya yi amfani da kayan aiki na yau da kullun a kansa, ba tare da wani abu mai sarkakiya ba.

Irin kwanukan lambobin farko biyar na NIST, tare da karin biyar daga ma'aunin ANSI X9.62, duk suna da tsari iri daya: lrand48() >> 23. Iri goma da ke raba dokar cirowa guda daya babbar shaida ce ta tsari guda daya a bayansu.

Sannan sai gwajin da ke raba dacewar sa'a daga ta gaske. Kowane yanayi mai bits 48 da aka dawo da shi yana sake samar da dukkan bits 160 na irinsa, abin da ba ya faruwa bisa kuskure. Dacewar wani bangare na iya zama karo na sa'a. Amma sake samar da kowane bit gaba daya abu ne mai matukar wuyar kawarwa, kuma shi ne dalilin da ya sa za a iya fadar sakamakon karara.

Gudanar da na'urar samarwa da baya don karanta agogon #

Dawo da yanayi yana gaya maka yadda aka samar da irin. Amma har yanzu bai gaya maka yaushe ba. Don haka, Claude ya yi amfani da wata siffa ta na'urar samarwar: tana tafiya da baya cikin sauki kamar yadda take tafiya gaba. Ga kowane iri, Claude ya mayar da yanayin baya mataki-mataki har sai da ya kai ga sigar da farawar srand48 za ta samar, tare da bits 16 na kasa daidai da 0x330E.

Wannan sigar sawun yatsa ce ta kanta. Idan za a iya mayar da yanayin da aka dawo da shi baya zuwa wani abu mai bits 16 na kasa a 0x330E, to an fara na'urar samarwar ne ta hanyar da aka saba, daga darajar iri da aka mika wa srand48. Kuma darajar irin, a wannan yanayin, lokaci ne. An samar da irin ne ta amfani da PRNG mai rauni wanda aka saita da Unix timestamp.

Tunanin cikin gida na Claude a wancan lokacin ya rubuta lokacin da ya gane abin: dukkan bits na sama sun fito fili (clear), bits 20, tare da yiwuwar sa'a ta 1e-6. Karamin bayani ne, amma yana nuna tunanin yana aiki a kan lambobin yayin da suke bayyana, ba a kan labarin da aka hada daga baya ba.

Kwanakin watan ba dukkansu suka fito sumul ba. Farawar kwanukan binary (binary-curve initialisations) guda hudu sun nuna Fabrairu 1984, wanda ba kwanan wata ba ne da wani zai yi tsammani ga na'urar samarwa da ake amfani da ita a karshen shekarun 1990. Da zarar an ninka lokutan (timestamps) sau biyu, farawar kwanukan binary sun nuna 17 ga Afrilu, 1998, makonni biyu bayan gudanar da P-224. Tushen bayanin ya kammala da cewa shirin na iya zama ya raba agogon gida biyu. Wannan an fada shi cikin taka-tsantsan, kuma ya kamata ya zama haka: bayani ne da aka ba da shawara kan me ya sa danyen darajojin suka nuna 1984, kuma ninkawar da ta sauka makonni biyu bayan wani gudanarwar tana nuna alama mai karfi ba tare da zama tabbaci na cikin shirin ba.

Ina ganin wannan shi ne bangare mafi gamsarwa na daukacin labarin, saboda wani dalili mai zaman kansa daga cryptography. Labarin karya da an kirkire shi zai samar da kwanakin wata masu tsari sumul. Wannan ya samar da wani mai banbarakwai, sannan gyaran da ya sanya kwanukan binary a cikin jeri mai ma'ana tare da gudanarwar P-224. Bayanai masu cakudewa wadanda ke warwarewa a karkashin zato mai sauki haka asali na gaske yake kama.

Maki na tushe (base points) ma suna da dabara #

Iri suna bayyana kwanukan. Maki na tushe (base points) batu ne daban, kuma tushen bayanin yana daukarsu daban.

Maki na tushe na kwanuka shida da aka samar a Mayu 1997 (P-192 da kwanukan misali biyar na ANSI X9.62 wato prime192v2, prime192v3 da prime239v1 zuwa v3) sun fito ne daga na'urar samar da lambobi masu bazuwa a cikin FreeLIP, kunshin lissafin manyan lambobi na Arjen Lenstra. Sigogin farko na dakin karatu na NTL na Victor Shoup sun zo da wannan lambar (code) iri daya. Wannan batu na biyu yana da muhimmanci don gwadawa: na'urar samarwar tana wanzuwa a matsayin lambar da aka wallafa wadda wasu za su iya karantawa, ba a matsayin kayan aiki na sirri ba.

Dabarar gajeriya ce. A saita na'urar samarwar da karamar lamba (small integer). A jefar da sakamako bakwai na farko. A samar da x-coordinates da ake gwadawa daga sakamako masu bits 26 da aka rage mod p. A maimaita wannan samarwar har sai x ya hau kan kwanar.

Claude ya gano irin na'urar samarwar: 20000 don P-192, 20420, 20507 da 20509 don prime239v1 zuwa v3, da kuma 20512 da 20517 don prime192v2 da v3. Wadannan kananan lambobi ne irin wadanda mutum ke bugawa da hannu a cikin shiri, wanda shi ne ainihin abin da matakin farko na dabarar ke nema.

Marubutan sun yi wani ikirari a nan mai saukin kai wanda ya cancanci a nakalto ma'anarsa: a iya saninsu, ba a taba bayyana wani maki na tushe na NIST ba kafin wannan. Taka-tsantsan din nasu ne, kuma ya dace. Ba sa ikirarin cewa sun bincika kowace takarda da aka taba rubutawa, sai dai iyakar iliminsu kawai.

Sannan dabarar ta daina aiki. Maki na tushe na P-224, P-256, P-384 da P-521 ba su dace da wannan dabarar ba, kuma marubutan ba su iya bayyana yadda aka zabe su ba. Rubutun da zai rufe wannan zai yi kama da mafi tsafta, amma marubutan sun bar shi a ciki. Bayanin ya rufe kwanuka shida daga Mayu 1997 kuma ya bar hudu daga cikin maki na tushe na kwanukan lambobin farko na NIST ba tare da bayani ba.

Wadanda aka gwada aka tsallake su #

Abin da aka gano na karshe shi ne wanda ya fi yiwuwar tayar da muhawara, don haka yana da kyau a karanta shi sosai.

Ga P-256, irin da aka wallafa shi ne na 1842 a jerin gwajin. Amma na 876 a cikin wannan gudanarwar shi ma yana ba da kwana mai darajar lambar farko (prime-order curve) mai girman da ya dace, tare da sigogin da aka zaba ta hanya daya kamar na kwanar da aka wallafa. A wasu kalmomin, aikin ya tsallake wata kwana wadda, a zahiri, za ta kasance mai karbuwa.

Wannan ba sau daya kawai ya faru ba. Gudanarwar P-224 ta tsallake irin wadannan kwanuka uku, kuma gudanarwa biyu da ke bayan kwanukan misali na ANSI X9.62 sun tsallake karin bakwai.

Marubutan sun fito fili game da abin da ba su sani ba. Ba su san wadanne sharudda aka yi amfani da su wajen tsallake kwanukan ba, ko kuma ko hakan wani sakamako ne da ba a nufa ba (unintentional artifact). Wannan iyaka ce, wadda aka fada a haka, kuma taka-tsantsan din a kan "sakamakon da ba a nufa ba" shi ne matsayi na gaskiya: bayanan sun nuna cewa an tsallake wasu da aka gwada amma ba su nuna me ya sa ba.

Daga nan sai suka ba da hukunci na tunani. Saboda adadin wadanda aka tsallake kadan ne, ba su yarda cewa abu ne mai yiwuwa a ce an zabi kwanuka masu rauni ba. Wannan tunani ne mai taka-tsantsan game da yiwuwa, kuma ya dogara ne akan kirga. Idan aiki ya tsallake 'yan kadan kawai a jimlace, akwai karamin fili kawai da za a iya neman rauni a cikinsa.

Gargadi daya ya rage, kuma tushen bayanin ya fada shi a matsayin yiwuwa ba abin da aka gano ba. Har yanzu ana iya tunanin cewa an zabi wadannan kwanukan ne don su sami wata siffa ta musamman a ka'ida ta hanyar bincike a kan lokuta (timestamps) da aka saka a na'urar samar da lambobi masu bazuwa. Wannan jita-jita ce ta tunani (speculation), kuma an yi mata alama a haka. Bai kamata a karanta ta a matsayin shaida ta komai ba, kuma bai kamata a yi shiru a jefar da ita ba. Rahoto na adalci yana ajiye ta a kan tebur a matsayin hanyar da ba a kore ta gaba daya ba, yayin da yake lura cewa marubutan suna daukar batun kwanuka masu rauni a matsayin abu marar yiwuwa.

Abin da sakamakon ya canza, da abin da ya bari a bude #

A fara da abin da bai canza ba. Kwanukan ba su karye ba. Tushen bayanin ya fadi haka sau biyu, a kalmomi daban-daban, kuma babu wani mataki da za a dauka. Duk wanda ke karanta wannan a matsayin dalilin cire P-256 ya karanta shi ba daidai ba.

Abin da yake canzawa shi ne matsayin wani tsohon kokwanto. Muddin ba a bayyana irin ba, cewa "NSA ce ta bayar da su ba tare da bayani ba" alamar tambaya ce da ke tsaye a kan ma'aunin. Wannan alamar tambayar yanzu tana da amsa ta zahiri. Irin ba hashes ba ne masu boyayyen ma'ana na wata jimlar sirri. Sakamako ne na lrand48 daga shirin da aka saita daga agogo, yana sake samar da kowane bit na kowane iri daga yanayin da aka dawo da shi. Bayanin da duk wanda ke da irin da aka wallafa zai iya gwadawa ya fi jita-jitar da babu wanda zai iya gwadawa.

Akwai darasi a nan ga wadanda suka kafa kamfanoni da injiniyoyi da ke gini a kan ma'aunin cryptography. Ikirarin asali yana da arhar yi amma yana da tsadar gwadawa, kuma gwajin da aka bayyana a nan ya dauki kasa da rabin sa'a da zarar an yi tambayar da ta dace. Na'urar duba, binciken 2^17 da ragewar lattice duk matakai ne na yau da kullun. Abin da aka rasa tsawon shekaru shi ne wani da zai yi aikin da ba shi da kwalliya na gwada su a kan lambobin da aka wallafa.

Zan kuma kauce wa jarabar zuzuta hanyar da aka bi. Tushen bayanin yana ba da rahoton sakamakon tawaga daya a kan takamaiman rukunin iri. Ba ya ikirarin cewa wannan hanyar za ta warware kowace lamba da ba a bayyana ba a cikin kowane ma'auni, kuma ba zan yi musu wannan ikirarin ba. Abin da yake nunawa ya fi kunkunta kuma, a nawa ganin, ya fi amfani: lokacin da asalin wata siga ya zama asiri, samfuri mai taka-tsantsan tare da zato mai ma'ana game da kayan aikin wancan zamanin zai iya gwada zaton cikin sauri kuma ya nuna aikinsa.

Aikin da ba a kammala ba ya fito fili sosai har zai iya zama ajandar bincike. Maki na tushe guda hudu, na P-224, P-256, P-384 da P-521, har yanzu ba su da bayani. Wadanda aka gwada aka tsallake, kamar 876 idan aka kwatanta da 1842 da aka wallafa don P-256, ba su da bayyanannen sharadi, kuma har yanzu babu wanda ya san ko tsallakewar da gangan ne ko kuma sakamako ne na yanayin aikin. Ko wadannan amsoshin suna raye a cikin tsoffin lambobin tushe (source code), a cikin tunanin abokan aiki da ke raye, ko kuma a cikin irin wannan farautar tsarin da ta warware irin, ita ce tambayar da ke bude. Irin sun dauki rabin sa'a; maki na tushe sun riga sun bijire wa wannan hanyar sau daya.

Tambayoyin da Ake Yawan Yi #

Shin wannan binciken yana karya kwanukan NIST ECDSA kamar P-256 ko P-384?

A'a. Gano cewa an samar da irin kwanukan ta hanyar lrand48 maimakon hash na kalmar sirri ba ya kawo wani karya tsarin sirri, sabon hari ko raunin lissafi a cikin kwanukan ECDSA, kuma ba a bukatar wani matakin gyara ko sauya tsari saboda wannan gano abin.

Ta yaya aka samar da irin kwanukan lambobin farko na NIST a zahiri?

An samar da kowane iri mai bits 160 ne ta hanyar kiran lrand48 linear congruential generator na dakin karatu na C (musamman lrand48() >> 23 don kwanukan lambobin farko biyar na NIST da kwanuka biyar na ANSI X9.62), wanda aka fara ta hanyar srand48 tare da Unix timestamp. Dawo da yanayin cikin gida mai bits 48 ta hanyar ragewar lattice ta Lenstra-Lenstra-Lovasz (LLL) yana sake samar da dukkan bits 160 na kowane irin da aka wallafa.

Me ya rage wanda ba a bayyana ba game da sigogin kwanar elliptic na NIST?

Yayin da aka gano cewa maki na tushe na P-192 da kwanuka biyar na ANSI X9.62 sun fito ne daga na'urar samar da lambobi masu bazuwa a cikin kunshin FreeLIP na Arjen Lenstra wanda aka saita da kananan lambobi, maki na tushe na P-224, P-256, P-384 da P-521 ba su dace da wannan dabarar ba, kuma sharuddan tsallake kwanukan farko masu darajar lambar farko (kamar na 876 kafin na 1842 a cikin P-256) sun ci gaba da kasancewa ba a san su ba.

Manazarta #

• Anthropic Frontier Red Team, (2026). Claude reveals ECDSA curve origins ⧉. Anthropic.
• National Institute of Standards and Technology, (2000). FIPS PUB 186-2: Digital Signature Standard (DSS) ⧉. NIST.
• Victor Shoup, (2024). NTL: A Library for doing Number Theory ⧉. libntl.org.

Bita ta ƙarshe 2026-11-01.

# Boyayyun iri (Seeds) da ke bayan kwanukan NIST ba Hashes ba ne tun fil azal, kuma Claude ya gano hakan a kasa da rabin sa'a — Sebastien Rousseau

> Originally published at [https://sebastienrousseau.com/ha/2026-10-26-boyayyun-iri-seeds-ke-bayan-kwanukan/](https://sebastienrousseau.com/ha/2026-10-26-boyayyun-iri-seeds-ke-bayan-kwanukan/)

Irin (seeds) da aka wallafa tare da kwanukan NIST P-192, P-224, P-256, P-384 da P-521 ba sakamakon hash ba ne tun farko. Sun fito ne daga na'urar samarwa ta lrand48 a dakin karatu na C, wanda aka saita da Unix timestamp. Claude ya gano...

Read the full article on sebastienrousseau.com: https://sebastienrousseau.com/ha/2026-10-26-boyayyun-iri-seeds-ke-bayan-kwanukan/

Boyayyun iri (Seeds) da ke bayan kwanukan NIST ba Hashes ba ne tun fil azal, kuma Claude ya gano hakan a kasa da rabin sa'a — Sebastien Rousseau

Irin (seeds) da aka wallafa tare da kwanukan NIST P-192, P-224, P-256, P-384 da P-521 ba sakamakon hash ba ne tun farko. Sun fito ne daga na'urar samarwa ta lrand48 a dakin karatu na C, wanda aka saita da Unix timestamp. Claude ya gano...

https://sebastienrousseau.com/ha/2026-10-26-boyayyun-iri-seeds-ke-bayan-kwanukan/

Boyayyun iri (Seeds) da ke bayan kwanukan NIST ba Hashes ba ne tun fil azal, kuma Claude ya gano hakan a kasa da rabin sa'a — Sebastien Rousseau

Irin (seeds) da aka wallafa tare da kwanukan NIST P-192, P-224, P-256, P-384 da P-521 ba sakamakon hash ba ne tun farko. Sun fito ne daga na'urar samarwa ta lrand48 a dakin karatu na C, wanda aka saita da Unix timestamp.

Ga abubuwan da ya kamata a lura da su na dabarun:

- Lambar kwanar da ta yi kama da an samo ta daga wani abu da kuma irin da babu wanda zai iya bayyanawa. NIST ta karbi ECDSA a shekarar 2000 a matsayin wani bangare na FIPS 186-2 Digital Signature Standard.
- Minti talatin, babu alama, da sakamakon da ba hash ba ne. A watan Oktoba na 2026, wani ma'aikacin Anthropic mai son sani ya tambayi wani samfurin bincike na cikin gida ko zai iya gano yadda aka samar da wadannan irin, ba tare da nuna alamar inda za a duba ba.
- Yadda na'urar duba ta gano sawun yatsan na'urar samarwa ta shekarun 1990. Hanyar da aka bi ta cancanci kulawa, domin ita ce bangaren da wasu mutane za su iya sake amfani da shi.
- Gudanar da na'urar samarwa da baya don karanta agogon. Dawo da yanayi yana gaya maka yadda aka samar da irin.

Menene hanyar ƙungiyar ku wajen magance ƙalubalen da aka kawo a wannan rubuce-rubucen?

→ https://sebastienrousseau.com/ha/2026-10-26-boyayyun-iri-seeds-ke-bayan-kwanukan/

#Asiri #Iri #Baya #Nist #Kwanuka

Sebastien Rousseau | CC-BY-4.0

@online{rousseau2026boyayyun,
  author  = {Rousseau, Sebastien},
  title   = {{Boyayyun iri (Seeds) da ke bayan kwanukan NIST ba Hashes ba ne tun fil azal, kuma Claude ya gano hakan a kasa da rabin sa'a — Sebastien Rousseau}},
  year    = {2026},
  url     = {https://sebastienrousseau.com/ha/2026-10-26-boyayyun-iri-seeds-ke-bayan-kwanukan/},
  urldate = {2026}
}

TY  - GEN
AU  - Rousseau, Sebastien
TI  - Boyayyun iri (Seeds) da ke bayan kwanukan NIST ba Hashes ba ne tun fil azal, kuma Claude ya gano hakan a kasa da rabin sa'a — Sebastien Rousseau
PY  - 2026
UR  - https://sebastienrousseau.com/ha/2026-10-26-boyayyun-iri-seeds-ke-bayan-kwanukan/
ER  -

Rousseau S. Boyayyun iri (Seeds) da ke bayan kwanukan NIST ba Hashes ba ne tun fil azal, kuma Claude ya gano hakan a kasa da rabin sa'a — Sebastien Rousseau. sebastienrousseau.com. 2026 Oct 26. Available from: https://sebastienrousseau.com/ha/2026-10-26-boyayyun-iri-seeds-ke-bayan-kwanukan/

Rousseau, Sebastien. "Boyayyun iri (Seeds) da ke bayan kwanukan NIST ba Hashes ba ne tun fil azal, kuma Claude ya gano hakan a kasa da rabin sa'a — Sebastien Rousseau." sebastienrousseau.com. October 26, 2026. https://sebastienrousseau.com/ha/2026-10-26-boyayyun-iri-seeds-ke-bayan-kwanukan/.

Rousseau, S. (2026, October 26). Boyayyun iri (Seeds) da ke bayan kwanukan NIST ba Hashes ba ne tun fil azal, kuma Claude ya gano hakan a kasa da rabin sa'a — Sebastien Rousseau. sebastienrousseau.com. https://sebastienrousseau.com/ha/2026-10-26-boyayyun-iri-seeds-ke-bayan-kwanukan/

Boyayyun iri (Seeds) da ke bayan kwanukan NIST ba Hashes ba ne tun fil azal, kuma Claude ya gano hakan a kasa da rabin sa'a — Sebastien Rousseau

Irin (seeds) da aka wallafa tare da kwanukan NIST P-192, P-224, P-256, P-384 da P-521 ba sakamakon hash ba ne tun farko. Sun fito ne daga na'urar samarwa ta lrand48 a dakin karatu na C, wanda aka saita da Unix timestamp. Claude ya gano...

Originally published at https://sebastienrousseau.com/ha/2026-10-26-boyayyun-iri-seeds-ke-bayan-kwanukan/ by Sebastien Rousseau.
Licensed under CC-BY-4.0.

SEBASTIEN ROUSSEAU · FOUNDER · ENGINEER