By · Last updated 2026-03-29

Rudi kwa BlogUsalama wa AI

Uvujaji Milioni 39 wa GitHub: Hatari ya AI ya Uandishi wa Msimbo

Asilimia 67 ya wasanidi programu wamewahi kwa bahati mbaya kufichua siri katika msimbo (GitGuardian 2025). Siri milioni 39 zilivuja kwenye GitHub mnamo 2024, ongezeko la asilimia 25 kwa mwaka.

March 29, 20268 dakika kusoma
GitHub secret leaksdeveloper AI securitycredential exposureMCP Server protectionGitGuardian 2025

Vitambulisho Milioni 39 Vilivyovuja kwa Mwaka Mmoja

Ripoti ya Octoverse 2024 ya GitHub iligundua siri milioni 39 zilizovuja kwenye GitHub mnamo 2024. Hiyo ni ongezeko la asilimia 25 kwa mwaka kutoka 2023. Siri hizo zinajumuisha funguo za API, nyuzi za hifadhidata, tokeni za uthibitisho, na vitambulisho vya wingu.

Sababu inajulikana. Wasanidi programu wanachangia msimbo wenye siri ndani. Siri zinatoka kwenye vikao vya utatuzi. Au zinawekwa kwa ugumu badala ya kuhifadhiwa katika vigeuzi vya mazingira. Kwa uvujaji milioni 39, hii si nadra. Ni kawaida.

Zana za AI Zinaongeza Njia ya Pili ya Uvujaji

Utafiti wa GitGuardian wa 2025 uligundua kwamba asilimia 67 ya wasanidi programu wamewahi kwa bahati mbaya kufichua siri katika msimbo. Tabia sawa zinazosababisha uvujaji wa GitHub pia husababisha uvujaji wa zana za AI.

Msanidi programu anabandika msimbo kwenye Claude, ChatGPT, au msaidizi mwingine wa AI kwa msaada. Msimbo huo mara nyingi una vitambulisho hai ndani yake. Modeli ya AI inapokea siri. Inaweza kuihifadhi katika historia ya mazungumzo. Inatumia kwa seva za mtoa huduma. Msanidi programu anapoteza udhibiti - bila onyo.

Mifano mitatu:

Utatuzi wa hifadhidata. Msanidi programu anabandika mkondo wa matatizo. Mkondo huo unajumuisha nyuzi ya muunganisho. AI inasoma nenosiri pia.

Ukaguzi wa mfumo. Msanidi programu anashiriki hati ya mfumo wa data. Hati hiyo ina funguo ya ufikiaji ya AWS na funguo ya siri. AI inapokea zote mbili.

Ukaguzi wa ujumuishaji wa API. Msanidi programu anaomba maoni juu ya ujumuishaji. Msimbo unajumuisha funguo ya API ya mshirika hai. Funguo inaondoka kwenye mtandao wa msanidi programu.

Katika kila kesi, lengo ni msaada halali. Uvujaji wa vitambulisho ni athari ya pembeni ya kumpa AI muundo wa kutosha. Hii ni mfumo sawa na uvujaji wa GitHub - si wa makusudi, ni wa kawaida tu.

Mifumo ya CI/CD Inakabiliwa na Hatari Sawa

Uvujaji wa siri za mfumo wa CI/CD uliongezeka kwa asilimia 34 mnamo 2024. Hati za ujenzi, usanidi wa usambazaji, na faili za miundombinu-kama-msimbo zote hupita kwenye ukaguzi wa AI sasa. Faili hizi mara nyingi zina vitambulisho vya wingu na tokeni za akaunti za huduma.

Zana za AI zinapofunika zaidi ya mzunguko wa maendeleo - ukaguzi, hati, utatuzi, uboreshaji - uso wa wazi hukua pamoja nazo.

Jinsi Muundo wa MCP Unavyozuia Uvujaji

Kwa timu zinazotumia Claude Desktop au Cursor IDE, muundo wa seva ya Model Context Protocol (MCP) unaweka kichujio cha vitambulisho katika njia kati ya msanidi programu na modeli ya AI.

Seva ya MCP inashughulikia kila maandishi yanayopita kwenye kikao. Msimbo uliobandiwa, mikondo ya matatizo, faili za usanidi, muundo wa utatuzi - yote hupita kwenye hatua ya kutoidhibitisha kabla modeli haijayaona.

Enjini inapata mifumo ya vitambulisho: miundo ya funguo ya API, nyuzi za hifadhidata, tokeni za OAuth, vichwa vya funguo za kibinafsi, na miundo ya kawaida timu yako ya usalama inavyofafanua. Kila mechi hubadilishwa na tokeni kabla ya usambazaji.

Hii inaonekana vipi kwa vitendo:

Msanidi programu anabandika mkondo wa matatizo wenye nyuzi ya muunganisho wa hifadhidata. Seva ya MCP inabadilisha nyuzi na [DB_CONNECTION_1]. AI inaona mkondo na tokeni badala yake. Inatoa msaada wa utatuzi kulingana na toleo la kutoidhibitishwa. Kitambulisho halisi hakijawahi kuondoka kwenye mtandao wa ndani.

Hii inazuia njia sawa ya uvujaji inayojaza GitHub na siri. Njia ni tofauti - zana za AI, si git commits - lakini suluhisho linafanya kazi sawa: izuie kabla haijatumwa.

Angalia muhtasari wetu wa usalama kwa jinsi anonym.legal inavyoshughulikia hili katika zana za AI na mifumo ya nyaraka, na kituo cha utiifu kwa udhibiti wa ukaguzi.

Ugunduzi Baada ya Ukweli ni Kuchelewa Sana

Baadhi ya timu hutumia uchanganuzi wa baada ya kuchangia kukamata siri zilizovuja. GitGuardian na truffleHog hufanya kazi vizuri kwa njia ya GitHub. Hawafuniki vikao vya zana za AI.

Siri inapofika kwenye seva za mtoa huduma wa AI, wazi imeisha. Uchanganuzi unaikuta baadaye. Kutoidhibitisha kwa safu ya MCP kunazuia isifike kwenye modeli kabisa.

Uvujaji milioni 39 wa GitHub unaandika njia moja. Wazi wa zana za AI ni tatizo sawa katika njia yenye ufuatiliaji mdogo na mlolongo wa ukaguzi. Kuzuia kabla ya usambazaji kunafunika vyote viwili.

Vyanzo

Tayari kulinda data yako?

Anza kuanonymisha PII na aina 285+ za vitu katika lugha 48.

About this page

We update this page when our platform or the law changes.

Read our founder note for how we work.

Each change shows up in the timestamp at the top.

Related reading

We follow these rules

  • GDPR (EU 2016/679).
  • ISO/IEC 27001:2022.
  • NIS2 (EU 2022/2555).
  • HIPAA safe harbor under 45 CFR § 164.514(b)(2).

Our promise

We do not sell your data.

We do not train models on your text.

We store your files in Germany.

You can delete your account at any time.

You own your work.

Where we run

Our servers live in Falkenstein, Germany.

We use Hetzner. They hold ISO 27001 certification.

All data stays in the EU.

Backups run every day.

Need help?

Email support@anonym.legal.

We reply within one business day.

How we test

We run a full check suite on every release.

Each surface gets its own sweep script and report.

Human reviewers spot-check the output each week.

We track recall and precision on a labelled set.

Bad runs block the deploy.

What we never do

  • We never sell your information to third parties.
  • We never train models on what you upload.
  • We never keep your work after you delete it.
  • We never share keys with any outside firm.
  • We never run ads inside the product.

Plans in plain words

We sell credits, not seats.

One credit covers one short job.

Long jobs use a few credits each.

You can top up at any time.

Unused credits roll over each month.

Read the plans page for current rates.

Who built this

A small team of engineers and lawyers built this.

We ship from Europe and work in the open.

Our founder note spells out why we started.

Where to start

How the parts fit

A browser add-on cleans text inside Chrome.

A Word plug-in handles drafts in Office.

A small desktop tool works on whole folders.

An agent protocol link feeds large models safely.

All four share one core engine and one rule set.

Words from our team

We started this work after a lunch about cookies.

One friend kept getting odd ads on her phone.

We asked why a court file leaked through a draft.

We sketched the first build on a napkin that week.

By month three we had a tiny demo for a friend.

She used it on her first case the next day.

Common questions we hear

Can the tool read scanned PDFs? Yes, with OCR.

Does it work on long files? Yes, in small chunks.

Can I roll my own rule set? Yes, save it as a preset.

Does it run offline? The desktop build runs offline.

Do you keep my files? No, the cloud build wipes after each run.

Will it learn from my work? No, we never train on inputs.

A short tour of the workflow

Upload a file or paste a snippet of prose.

Pick the entities you want gone from the draft.

Choose a method: replace, mask, hash, encrypt, or redact.

Press run and watch the side panel show each hit.

Skim the result and tweak any rule that misfired.

Save the cleaned file or send it to a teammate.