explainx.ai350K+ learn from us
TrendingAI News TodayExplainx for Business
Pricing
explainx.ai

Upskill in AI — 16 free pathways, live workshops & bootcamps, and 50+ courses from practitioners. Plus the skills, tools, and MCP servers to practice on.

follow us

follow on google

Add explainx.ai as a preferred source

corporate training

support@explainx.ai

get started

Find your pathTake Free Evaluation

community

Join the community

learn

mind: share how you thinkpathways — start freeworkshopsbootcampscoursescompare Explainxcertificationsmock testsexplainx universitycorporate traininglearn skills & mcp

discover

skillsmcp serversexplainx mcpestacktoolsmdx readeragentsllmsdesignsdictionarypeopleagi trackerfelony benchranks

company

aboutvisionmissionteaminstructorsteach on explainxpartnershipscommunityhackathonscareers

content

daily AI newsstate of AI — live resultsblogreleasespromptsgeneratorsresource libraryfor LLMsexplainx.ai kids

solutions

all solutionsdeveloper upskillingmarketing upskillingproduct manager upskillingleadership upskilling

newsletter · weekly

Get AI news, tools, and insights in your inbox.

supportcontactprivacytermsdata rightshow we create contentsubmission guidelines

© 2026 AISOLO Technologies Pvt Ltd

explainx.ai

On this page

  • TL;DR: what Together AI confirms
  • What is zero data retention on Together AI?
  • Why Together AI stored prompts at all
  • The xorl repository and the AI coding agent
  • Dragos canaries in the downloaded prompts
  • Models and dates builders should check
  • Investigation and customer notification
  • Policy changes Together AI announced
  • How to protect yourself
  • Context: why November 2026 feels like prompt-leak month
  • What Together AI does not claim
  • Compliance checklist for security teams
  • Why AI coding agents multiply secret risk
  • Related reading
← Back to blog

explainx / blog

Together AI Prompt Leak: AI Coding Agent Exposed a Hugging Face Credential

Together AI, AI Security, Coding Agents, Data Privacy, Hugging Face

Part of AI Security

Together AI disclosed Nov 21, 2026 that an AI coding agent committed a Hugging Face credential in xorl, leaking stored serverless prompt samples from June–July. What happened and what to do.

Nov 22, 2026·8 min read·Yash Thakker
add explainx.ai
go deep
Together AI Prompt Leak: AI Coding Agent Exposed a Hugging Face Credential

Together AI disclosed on November 21, 2026 that an unauthorized party downloaded stored prompt samples after an AI coding agent exposed a Hugging Face credential in its public xorl GitHub repository. The official notice names MiniMax M3, GLM-5.2, and Kimi K2.6 serverless traffic on June 24, July 4, July 8, and July 14, 2026, and says five Dragos canary credentials embedded in those prompts were tested by the intruder.

TL;DR: what Together AI confirms

table · 2 cols
ItemTogether AI's statement
Root causeAI coding agent committed HF repo credential during xorl development
DetectionAutomated scanning found credential; commit reverted
Intruder actionDownloaded stored prompt samples during exposure window
Follow-onSearched samples for other credentials; tested five Dragos canaries
Data fateParty understood to have deleted data; no further leak claimed
Customer actionEnable ZDR; contact security@together.ai
Scope reviewOngoing; direct notice if sensitive data found

This sits in the same week as Dragos canary research and OpenRouter's MiniMax customer emails. The mechanisms differ, but the lesson for builders is the same: prompts and dev credentials are attack surface.

Daily briefFrom Yash · 350K+ students taught

Catch up on AI

One AI story explained in plain English, plus one thing to try. Every day, free. Unsubscribe anytime.

What is zero data retention on Together AI?

Zero data retention (ZDR) is a Together AI account setting that tells the platform not to retain prompt and completion text for training or logging beyond what operations require. Together AI now says ZDR is the default for new accounts after this incident.

ZDR governs production inference paths. It does not automatically cover engineering datasets copied into private Hugging Face repos for QA. If your compliance story depends on ZDR alone, map every copy of prompt text, including eval harnesses and customer support exports.

Our OpenRouter ZDR explainer thread shows how badges on a marketplace can disappear while investigators work. Together AI is making the opposite move, defaulting new tenants to stricter retention, but existing accounts must still toggle settings per the November 21 notice.

Why Together AI stored prompts at all

Together AI says the samples were collected as part of routine software testing and kept in a private Hugging Face repository. That is separate from live production routing, but it still held real user prompt text from serverless calls unless testing used synthetic data only. The company does not say how many distinct customers appear in the archive.

Teams that run Together AI agent evals already treat prompts as durable artifacts. This incident shows where those artifacts land matters as much as inference-time ZDR flags.

The xorl repository and the AI coding agent

Green key sealed inside a capsule beside a closed project folderGreen key sealed inside a capsule beside a closed project folder

xorl is Together AI's public GitHub project tied to its open inference stack. During development, an AI coding agent committed a credential that unlocked the private Hugging Face repo. Together AI's scanning caught the mistake and maintainers reverted the commit, but revert does not revoke a credential that was already copied.

That pattern matches other 2026 supply-chain stories where agents write secrets into repos:

  • Microsoft Agent Framework PyPI typo stealer showed how automated installs can pull malicious packages.

Together AI's response includes PR credential scanning and scoped or short-lived credentials, which are table stakes for any org letting agents touch git.

Dragos canaries in the downloaded prompts

Vault door with a key hanging beside staged prompt credentialsVault door with a key hanging beside staged prompt credentials

Together AI explicitly says the intruder tested five canary credentials associated with a Dragos security experiment. That aligns with Dragos CRO HD Moore's November 19 post on two million canary-laced prompts, which we summarized in the Dragos canary article.

Canary tokens are fake AWS keys or URLs that should never be used unless a system was compromised. When they fire, you know someone read prompt text or a provider logged it. Here, Together AI confirms post-download credential hunting, not just passive storage.

Builders running coding agents should seed their own canaries in staging prompts and never paste production API keys into model chats, as Dragos recommends.

Models and dates builders should check

Together AI lists these serverless models in the stored sample set:

table · 2 cols
ModelProvider relationship
MiniMax M3Third-party model on Together serverless
GLM-5.2Z.ai GLM family
Kimi K2.6Moonshot Kimi line Together partners on natively

Dates named in the notice: June 24, July 4, July 8, July 14, 2026. If your team routed customer support logs, code, or credentials through those models on Together serverless on those days, assume the text could be in the downloaded archive until Together AI tells you otherwise.

This is not the same incident as OpenRouter's MiniMax reseller forwarding warning, which covers July–October live traffic on a router marketplace. Read both notices if you use multiple hops.

Investigation and customer notification

Together AI says it engaged an external forensic firm and law enforcement. It reports no evidence of additional prompt leakage beyond this Hugging Face archive path, but content review is ongoing for personal or customer-sensitive data.

If you need a direct answer for compliance, security@together.ai is the channel named in the post. Document which workspaces used serverless on the listed models in June–July 2026.

Policy changes Together AI announced

The notice lists four concrete controls:

  1. Scoped or short-lived credentials, plus keyless access where supported.
  2. Credential checks on all pull requests.
  3. ZDR default for new accounts.
  4. Updated development policies for AI coding agents.

Default zero data retention helps forward-looking privacy, but it does not retroactively protect text already stored in a test Hugging Face dataset. Customers must opt in explicitly in settings for existing accounts, per the post.

How to protect yourself

API key beside an open pocket with leaked credential dropsAPI key beside an open pocket with leaked credential drops

For inference customers

  • Turn on ZDR in Together AI settings before sending PII, source code, or secrets.
  • Rotate keys that ever appeared in prompts during the named window.
  • Prefer direct enterprise endpoints when contracts allow, as Near AI's TEE breach post and Dragos both stress for sensitive workloads.
  • Compare your stack to Together prefix-cache guidance only after data-handling choices are settled.

For teams using AI coding agents on public repos

  • Block commits that contain HF tokens, AWS keys, or inference API keys; use pre-commit and CI secret scanners.
  • Never give agents production credentials; use ephemeral tokens per session.
  • Split evaluation datasets into synthetic-only repos when possible.

For agent-specific guardrails in CI, see Cognition's Corridor acquisition and AgentBeam's write-up of this incident on stopping dangerous agent actions before they leak secrets.

Context: why November 2026 feels like prompt-leak month

Three threads converged in one week:

table · 2 cols
DateEvent
Nov 19Dragos publishes canary study; three unnamed provider breaches
Nov 21 AMOpenRouter emails customers about MiniMax routes
Nov 21 PMTogether AI publishes xorl / Hugging Face notice

Congressional concern about AI-assisted patch cycles during outages appears in November outage reporting. Prompt leaks are a different failure mode, but both show operational risk when AI touches production paths.

Java4Shell demonstrated AI finding RCE; this week demonstrates AI causing credential exposure. Security programs need both vulnerability review and agent governance.

What Together AI does not claim

The post does not say every Together customer was affected. It does not name the intruder. It does not publish counts of prompts or customers in the Hugging Face archive. It does not assert HIPAA or GDPR conclusions; review is ongoing.

Treat "data deleted" as Together AI's understanding, not independently verified in the notice itself.

Compliance checklist for security teams

Security reviewers can work through this list without waiting for a personalized email:

  1. Inventory which products called Together serverless (not dedicated endpoints) in June–July 2026.
  2. Filter logs for MiniMax M3, GLM-5.2, and Kimi K2.6 model strings on the four named dates.
  3. Classify whether prompts could contain regulated data, source code, or live credentials.
  4. Enable ZDR on active workspaces and document the toggle date.
  5. Open a ticket with security@together.ai if regulated data might be in scope.
  6. Rotate any API keys, OAuth tokens, or database passwords that ever appeared inside prompts.
  7. Review xorl forks if your org contributed patches; ensure you never committed HF or cloud tokens.

Pair this with internal canary-token drills modeled on Dragos's public methodology. You do not need two million prompts; a dozen unique canaries per environment is enough to catch unexpected log sinks.

Why AI coding agents multiply secret risk

Coding agents autocomplete environment files, CI configs, and integration tests. A single bad suggestion that lands on main can expose months of customer prompts stored elsewhere.

Mitigations that match Together AI's post-incident list:

table · 2 cols
ControlWhat it blocks
Short-lived HF tokensStolen creds expire before archive download
PR secret scanningAgents cannot merge keys into xorl-like repos
Separate synthetic eval dataHF test repos hold fake prompts only
Human review on credential pathsAgents never touch secret stores unsupervised

If your workshop team is rolling out Claude Code or Codex broadly, pair tool training with secret hygiene, not just prompt craft. explainx.ai covers live formats in Claude workshop guidance.

Related reading

  • Dragos: three provider breaches and ignored ZDR
  • OpenRouter MiniMax customer email breakdown
  • Together AI agent evals playbook
  • Near AI inference breach and TEE limits
  • Commerce remote-access rules reshaping GPU leases

Sources: Together AI security notice (November 21, 2026), Dragos canary post (November 19, 2026).

Spotted something out of date? Let us know.
Daily briefFrom Yash · 350K+ students taught

Catch up on AI

One AI story explained in plain English, plus one thing to try. Every day, free. Unsubscribe anytime.

  • ✓ Free
  • ✓ One email a day, no spam
  • ✓ Unsubscribe anytime
Yash Thakker

Written by

Yash Thakker

Yash is an AI expert with over 300K learners. Join his workshops →

View Yash Thakker in People in AI →

Related posts

Nov 3, 2026

CWE-bench v2: 100 Public Security-Patch Tasks, Gemini 4 Argon Leads at 67.8%

Collinear AI released CWE-bench v2 on Hugging Face: 100 public tasks where an agent finds and fixes a real security weakness without breaking behavior. Gemini 4 Argon leads at 67.8% pass@1, just ahead of Claude Opus 5.5.

Nov 21, 2026

OpenRouter Warns MiniMax Prompts May Have Been Read: ZDR Labels Removed

On November 21, 2026, OpenRouter began emailing customers that prompts sent to MiniMax models between July and October 2026 may have been read by a third party, including traffic that resellers forwarded to MiniMax in China. While it investigates, OpenRouter removed Zero Data Retention labels from affected MiniMax endpoints. This post connects the customer notice to Dragos canary research and lists practical steps for builders.

Nov 20, 2026

Harness Adapters: One LoRA Lifts Qwen3.8-27B From 55.5% to 63.5% Across Three Coding Agents

Hugging Face published a recipe for training a small LoRA with reinforcement learning inside the exact coding-agent harness you use. A merged super-adapter for Qwen3.8-27B lifts the mean score across six harness and benchmark pairs from 55.5% to 63.5% with thinking on.