# principle of ephemerality

A software-security principle that recommends making credentials, keys, signatures, attestations, and infrastructure ephemeral to reduce risk.

This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles.

## Who collects the keys when an AI agent retires?

DevFeed: [Who collects the keys when an AI agent retires?](<https://devfeed.tech/articles/who-collects-the-keys-when-an-ai-agent-retires-129962.md>)

Original publisher: [Read original article](<https://www.helpnetsecurity.com/2026/11/12/sprinto-ai-agent-offboarding/>)

Author: Help Net Security

Published: 2026-11-12T06:00:57Z

Content type: opinion

Language: en

Sources: [Help Net Security](<https://devfeed.tech/sources/help-net-security.md>)

Topics: [AI Agent](<https://devfeed.tech/topics/ai-agent.md>), [Access Control](<https://devfeed.tech/topics/access-control.md>), [Secrets Management](<https://devfeed.tech/topics/secrets-management.md>), [principle of ephemerality](<https://devfeed.tech/topics/principle-of-ephemerality.md>), [human review](<https://devfeed.tech/topics/human-review.md>)

Tags: [access-management](<https://devfeed.tech/tags/access-management.md>), [agent](<https://devfeed.tech/tags/agent.md>), [agentic-ai](<https://devfeed.tech/tags/agentic-ai.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agent](<https://devfeed.tech/tags/ai-agent.md>), [audit](<https://devfeed.tech/tags/audit.md>), [compliance](<https://devfeed.tech/tags/compliance.md>), [credentials](<https://devfeed.tech/tags/credentials.md>), [cybersecurity](<https://devfeed.tech/tags/cybersecurity.md>), [decision](<https://devfeed.tech/tags/decision.md>), [don-t-miss](<https://devfeed.tech/tags/don-t-miss.md>), [expert-analysis](<https://devfeed.tech/tags/expert-analysis.md>), [hot-stuff](<https://devfeed.tech/tags/hot-stuff.md>), [it](<https://devfeed.tech/tags/it.md>), [keys](<https://devfeed.tech/tags/keys.md>), [net](<https://devfeed.tech/tags/net.md>), [news](<https://devfeed.tech/tags/news.md>), [opinion](<https://devfeed.tech/tags/opinion.md>), [records](<https://devfeed.tech/tags/records.md>)

### AI overview

AI agent retirement requires more than deleting an entry from a dashboard: credentials, delegated work and evidence of past decisions can persist. The article recommends issuing short-lived, narrowly scoped credentials where possible, tracing and testing remaining access after shutdown, and locating scheduled or delegated work so it can be reassigned or stopped. Preserve decision-time context--including model and prompt versions, input data, tool calls and applicable policy--because a final configuration snapshot cannot explain earlier decisions. Keep a retirement record with timestamps for identity disablement and the last action under that identity, plus the agent's owner and decision-evidence location; decide which cached or vector-store data to retain or delete. The article describes Sprinto's registry, owner sign-off, data decision and logged offboarding as one implementation. It argues that a system controlling access and maintaining versioned behavior histories becomes necessary as agent numbers grow.

### Source excerpt

Think about the last time a good engineer left your company. Someone collected the laptop and the badge. IT revoked their logins before the end of the day. Their manager reassigned the open tickets, and HR filed their records somewhere they'll stay for years, in case anyone ever challenges a decision they made. Now think about the last time an AI agent left. Someone probably deleted it from the dashboard. I've started describing agents as ... More -> The post Who collects the keys when an AI agent retires? appeared first on Help Net Security.

## Hackers stole AI credentials to get $740,000 in free compute; Railway picked up the bill

DevFeed: [Hackers stole AI credentials to get $740,000 in free compute; Railway picked up the bill](<https://devfeed.tech/articles/hackers-stole-ai-credentials-to-get-740-000-in-free-compute-railway-picked-up-the-bill-129744.md>)

Original publisher: [Read original article](<https://www.computerworld.com/article/4247584/hackers-stole-ai-credentials-to-get-740000-in-free-compute-railway-picked-up-the-bill-2.html>)

Author: Taryn Plumb

Published: 2026-11-12T02:56:11Z

Content type: news

Language: en

Sources: [Computerworld](<https://devfeed.tech/sources/computerworld.md>)

Topics: [credential-based attacks](<https://devfeed.tech/topics/credential-based-attacks.md>), [Secrets Management](<https://devfeed.tech/topics/secrets-management.md>), [ai security](<https://devfeed.tech/topics/ai-security.md>), [Security](<https://devfeed.tech/topics/security.md>), [principle of ephemerality](<https://devfeed.tech/topics/principle-of-ephemerality.md>), [hosting](<https://devfeed.tech/topics/hosting.md>), [AI Infrastructure](<https://devfeed.tech/topics/ai-infrastructure.md>)

Tags: [22](<https://devfeed.tech/tags/22.md>), [about](<https://devfeed.tech/tags/about.md>), [agent](<https://devfeed.tech/tags/agent.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-model](<https://devfeed.tech/tags/ai-model.md>), [ai-model-training](<https://devfeed.tech/tags/ai-model-training.md>), [ai-usage](<https://devfeed.tech/tags/ai-usage.md>), [allowed](<https://devfeed.tech/tags/allowed.md>), [also](<https://devfeed.tech/tags/also.md>), [artificial-intelligence](<https://devfeed.tech/tags/artificial-intelligence.md>), [artificial-intelligence-cyberattacks-cybercrime-hacking-security](<https://devfeed.tech/tags/artificial-intelligence-cyberattacks-cybercrime-hacking-security.md>), [attack](<https://devfeed.tech/tags/attack.md>), [bill](<https://devfeed.tech/tags/bill.md>), [build](<https://devfeed.tech/tags/build.md>), [but](<https://devfeed.tech/tags/but.md>), [card](<https://devfeed.tech/tags/card.md>), [compute](<https://devfeed.tech/tags/compute.md>), [credentials](<https://devfeed.tech/tags/credentials.md>), [cyberattacks](<https://devfeed.tech/tags/cyberattacks.md>), [cybercrime](<https://devfeed.tech/tags/cybercrime.md>), [edition-us](<https://devfeed.tech/tags/edition-us.md>), [free](<https://devfeed.tech/tags/free.md>), [hackers](<https://devfeed.tech/tags/hackers.md>), [hacking](<https://devfeed.tech/tags/hacking.md>), [language-en](<https://devfeed.tech/tags/language-en.md>), [railway](<https://devfeed.tech/tags/railway.md>), [security](<https://devfeed.tech/tags/security.md>), [storytype-news](<https://devfeed.tech/tags/storytype-news.md>), [up](<https://devfeed.tech/tags/up.md>)

### AI overview

Attackers exploited two Railway infrastructure weaknesses, accessed stored build histories and source code, then obtained credentials that could decrypt customer environment variables. Over nearly 26 hours, they accessed a subset of roughly 100,000 variables--at most 1.85% of workspaces--and generated about $740,000 in fraudulent AI charges. Railway says it evicted the attackers, reset linked-account tokens, worked with providers to revoke potentially compromised keys, and is reimbursing customers. It advises rotating sensitive variables and keys, reviewing account and registry audit logs, and removing public database access. The article also recommends checking AI usage and bills, alerting on unfamiliar traffic, and limiting key permissions and spending. Railway says it is moving to short-lived credentials and logging each decryption; the investigation was ongoing, and the excerpt says the company could not yet detail the exploits' full scope.

### Source excerpt

As AI model training and distillation increase, AI tokens are becoming a piping hot currency for threat actors. Cloud hosting provider Railway found out the hard way; it has confirmed a near 26-hour battle against attackers targeting AI provider credentials to sell to AI 'transfer stations' for use in model distillation. The crooks generated about $740,000 in fraudulent AI usage charges using the stolen credentials and potentially had access to up to 100,000 workspace credentials. Railway said the attacker's database searches looked for AI provider keys by name before anything else; some queries also targeted cloud, database, and payment service credentials. The company disclosed the incident on October 22, and posted a follow-up report this week. Because of the ongoing investigation, it is not able to share the specific details and scale of the exploits, but advised users to revoke and reapply sensitive environmental variables stored on Railway prior to October 21. "AI credentials are becoming the new credit card numbers," said David Shipley of Beauceron Security. "So-called LLMjacking is expected to displace Cryptojacking as a cheap, scalable way to make money." A 26-hour battle The attack began on October 20, when hackers exploited a previously unknown vulnerability in Railway's build infrastructure. They were able to evade filesystem isolation protections and obtain an internal credential that they used to access stored build histories. From there, they retrieved a copy of Railway's internal source code and analyzed it to identify additional avenues of compromise. They then exploited another vulnerability involving an improperly configured trust relationship that allowed them to gain even more access to Railway's infrastructure and obtain credentials used to decrypt customer environment variables. In a nearly 26-hour battle, Railway's engineering team worked 24-7 in shifts and the escalating situation became a "contest of agent speed," the company explained. The

## What October's security incidents taught us about exposed credentials

DevFeed: [What October's security incidents taught us about exposed credentials](<https://devfeed.tech/articles/what-october-s-security-incidents-taught-us-about-exposed-credentials-118997.md>)

Original publisher: [Read original article](<https://1password.com/blog/october-security-incidents-exposed-credentials>)

Author: info@1password.com (Mariane Bekker)

Published: 2026-11-04T00:00:00Z

Content type: tutorial

Language: en

Sources: [1Password](<https://devfeed.tech/sources/blog-on-1password-blog.md>)

Topics: [Secrets Management](<https://devfeed.tech/topics/secrets-management.md>), [Security](<https://devfeed.tech/topics/security.md>), [credential-based attacks](<https://devfeed.tech/topics/credential-based-attacks.md>), [principle of ephemerality](<https://devfeed.tech/topics/principle-of-ephemerality.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [ai-api-keys](<https://devfeed.tech/tags/ai-api-keys.md>), [ai-developer](<https://devfeed.tech/tags/ai-developer.md>), [credentials](<https://devfeed.tech/tags/credentials.md>), [developers](<https://devfeed.tech/tags/developers.md>), [environment-variables](<https://devfeed.tech/tags/environment-variables.md>), [exposed](<https://devfeed.tech/tags/exposed.md>), [for-developers](<https://devfeed.tech/tags/for-developers.md>), [incidents](<https://devfeed.tech/tags/incidents.md>), [infrastructure](<https://devfeed.tech/tags/infrastructure.md>), [october](<https://devfeed.tech/tags/october.md>), [permissions](<https://devfeed.tech/tags/permissions.md>), [pipeline](<https://devfeed.tech/tags/pipeline.md>), [secrets](<https://devfeed.tech/tags/secrets.md>), [security](<https://devfeed.tech/tags/security.md>), [security-incidents](<https://devfeed.tech/tags/security-incidents.md>), [workflows](<https://devfeed.tech/tags/workflows.md>)

### AI overview

October's reported security incidents show how credentials can expose customer secrets and provide paths into additional systems. Reports described internal credentials reaching infrastructure, databases, and environment variables; long-lived plaintext admin credentials; exposed encrypted secrets; unauthorized API-key use; and a public key that remained usable after its file was deleted, later contributing to a release-pipeline compromise. The reports do not establish a single attacker or vulnerability behind the incidents. Recommended steps are to reduce plaintext copies by using secret references or managed storage, limit workflow permissions and separate development from production credentials, use SSH agents to reduce private-key copies, understand which component receives real credentials when giving agents access, and revoke or rotate potentially leaked credentials with the issuing provider. Also review logs and update connected systems: deleting a file does not invalidate an existing key, and managed storage cannot prevent a compromised application from exposing credentials it receives.

### Source excerpt

October brought a series of security incidents across AI and developer platforms. Several platforms reported incidents involving credentials or environment variables. Across several of those disclosures, access to one system became a way to reach credentials for others. For developers, the lesson is practical: we need to manage credentials throughout the workflows that use them. That means reducing plaintext copies, limiting permissions, and making it easier to revoke access when something goes wrong. What happened across October An October 30 industry report initially listed nine incidents affecting AI developer platforms, with more added in November. Across that roundup and vendor disclosures, several patterns stood out. Internal credentials exposed customer secrets.Reports described internal credentials being used to access infrastructure, platform databases, and environment variables. Another preliminary report identified long-lived admin credentials stored in plaintext and usable from the public internet as a root cause. Customers were urged to rotate stored credentials. One platform recommended rotating secrets and updating credentials for connected services after identifying unauthorized credential access. Another reported exposure of encrypted secrets and potential decryption of some environment variables. Deleting a leaked key didn't invalidate it. An admin key was committed to a public repository. The file was removed, but the key wasn't revoked. Months later, an attacker used it in an attack that ultimately compromised a release pipeline. Unauthorized API usage. Customers were urged to assume environment variables were compromised and rotate secrets. It also reported that customers had observed unauthorized use of AI API keys supplied to the platform. These incidents had different entry points and impacts. The public reports don't establish one attacker or vulnerability behind the entire wave, but they do show credentials serving both as a target and as a

## Cisco Rethinks Privileged Infrastructure Access

DevFeed: [Cisco Rethinks Privileged Infrastructure Access](<https://devfeed.tech/articles/cisco-rethinks-privileged-infrastructure-access-100834.md>)

Original publisher: [Read original article](<https://blogs.cisco.com/security/cisco-rethinks-privileged-infrastructure-access>)

Author: Matt Caulfield

Published: 2026-10-26T20:00:16Z

Content type: article

Language: en

Sources: [Cisco Blogs](<https://devfeed.tech/sources/cisco-blogs.md>)

Topics: [identity and access management](<https://devfeed.tech/topics/identity-and-access-management.md>), [Access Control](<https://devfeed.tech/topics/access-control.md>), [least privilege](<https://devfeed.tech/topics/least-privilege.md>), [principle of ephemerality](<https://devfeed.tech/topics/principle-of-ephemerality.md>)

Tags: [access-management](<https://devfeed.tech/tags/access-management.md>), [accountability](<https://devfeed.tech/tags/accountability.md>), [actor](<https://devfeed.tech/tags/actor.md>), [administrator](<https://devfeed.tech/tags/administrator.md>), [advanced](<https://devfeed.tech/tags/advanced.md>), [agent](<https://devfeed.tech/tags/agent.md>), [agentic](<https://devfeed.tech/tags/agentic.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai-cybersecurity](<https://devfeed.tech/tags/ai-cybersecurity.md>), [cisco](<https://devfeed.tech/tags/cisco.md>), [cisco-privileged-infrastructure-access](<https://devfeed.tech/tags/cisco-privileged-infrastructure-access.md>), [critical](<https://devfeed.tech/tags/critical.md>), [cryptographic](<https://devfeed.tech/tags/cryptographic.md>), [identity](<https://devfeed.tech/tags/identity.md>), [identity-security](<https://devfeed.tech/tags/identity-security.md>), [infrastructure](<https://devfeed.tech/tags/infrastructure.md>), [infrastructure-access](<https://devfeed.tech/tags/infrastructure-access.md>), [keep](<https://devfeed.tech/tags/keep.md>), [moving](<https://devfeed.tech/tags/moving.md>), [network-security](<https://devfeed.tech/tags/network-security.md>), [privileged-infrastructure-access](<https://devfeed.tech/tags/privileged-infrastructure-access.md>), [security](<https://devfeed.tech/tags/security.md>), [with](<https://devfeed.tech/tags/with.md>), [work](<https://devfeed.tech/tags/work.md>)

### AI overview

Cisco Privileged Infrastructure Access replaces standing infrastructure privileges with on-demand, identity-bound sessions. The article says cryptographic identity and short-lived, least-privilege access let administrators and agents reach devices for a task, with privilege expiring when the work ends and session activity recorded. For customers using Cisco ISE and TACACS+, it can use existing policy and device context to govern access to Cisco routers and switches; the model can also extend to servers, Kubernetes clusters, and cloud resources. People, AI agents, and automation scripts use task-specific controls, with activity attributed to the actor. The article presents this as a faster governed path that reduces reliance on static credentials, shared accounts, and workarounds; it does not provide independent performance or security results.

### Source excerpt

Cisco Privileged Infrastructure Access replaces standing admin access with identity-bound, cryptographic controls that keep critical work moving.

## Cisco Rethinks Privileged Infrastructure Access

DevFeed: [Cisco Rethinks Privileged Infrastructure Access](<https://devfeed.tech/articles/cisco-rethinks-privileged-infrastructure-access-102531.md>)

Original publisher: [Read original article](<https://blogs.cisco.com/security/cisco-rethinks-privileged-infrastructure-access/>)

Author: Matt Caulfield

Published: 2026-10-26T20:00:16Z

Content type: release

Language: en

Sources: [Security @ Cisco](<https://devfeed.tech/sources/security-cisco-blogs.md>)

Topics: [Cisco](<https://devfeed.tech/topics/cisco.md>), [least privilege](<https://devfeed.tech/topics/least-privilege.md>), [identity and access management](<https://devfeed.tech/topics/identity-and-access-management.md>), [Access Control](<https://devfeed.tech/topics/access-control.md>), [principle of ephemerality](<https://devfeed.tech/topics/principle-of-ephemerality.md>)

Tags: [access-management](<https://devfeed.tech/tags/access-management.md>), [accountability](<https://devfeed.tech/tags/accountability.md>), [actor](<https://devfeed.tech/tags/actor.md>), [admin](<https://devfeed.tech/tags/admin.md>), [administrator](<https://devfeed.tech/tags/administrator.md>), [agent](<https://devfeed.tech/tags/agent.md>), [agentic](<https://devfeed.tech/tags/agentic.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-cybersecurity](<https://devfeed.tech/tags/ai-cybersecurity.md>), [cisco](<https://devfeed.tech/tags/cisco.md>), [cisco-privileged-infrastructure-access](<https://devfeed.tech/tags/cisco-privileged-infrastructure-access.md>), [critical](<https://devfeed.tech/tags/critical.md>), [cryptographic](<https://devfeed.tech/tags/cryptographic.md>), [identity](<https://devfeed.tech/tags/identity.md>), [identity-security](<https://devfeed.tech/tags/identity-security.md>), [infrastructure](<https://devfeed.tech/tags/infrastructure.md>), [infrastructure-access](<https://devfeed.tech/tags/infrastructure-access.md>), [keep](<https://devfeed.tech/tags/keep.md>), [moving](<https://devfeed.tech/tags/moving.md>), [network-security](<https://devfeed.tech/tags/network-security.md>), [privileged-infrastructure-access](<https://devfeed.tech/tags/privileged-infrastructure-access.md>), [security](<https://devfeed.tech/tags/security.md>), [work](<https://devfeed.tech/tags/work.md>)

### AI overview

Cisco Privileged Infrastructure Access is presented as an on-demand alternative to standing infrastructure credentials. It uses cryptographic identity and short-lived, least-privilege sessions that expire when a task ends, with session activity recorded to show who accessed what under which policy and what changed. Cisco says the approach is intended to provide network administrators and agents a direct, consistent path to critical devices while reducing reliance on shared secrets and broad persistent access. For customers using Cisco ISE and TACACS+, it can use existing policy and device context to govern access to Cisco routers and switches; the same model can extend to servers, Kubernetes clusters, and cloud resources. Humans, delegated AI agents, and scripts receive task-specific controls with activity attributed to the actor. The source describes the product's intended benefits and scope but does not provide independent validation or implementation details.

### Source excerpt

Cisco Privileged Infrastructure Access replaces standing admin access with identity-bound, cryptographic controls that keep critical work moving.

## Vercel moved to kill static credentials. Then developers spotted its own deployment token.

DevFeed: [Vercel moved to kill static credentials. Then developers spotted its own deployment token.](<https://devfeed.tech/articles/vercel-moved-to-kill-static-credentials-then-developers-spotted-its-own-deployment-token-98409.md>)

Original publisher: [Read original article](<https://thenewstack.io/vercel-oidc-static-credentials/>)

Author: Amanda Caswell

Published: 2026-10-23T20:04:16Z

Content type: news

Language: en

Sources: [The New Stack](<https://devfeed.tech/sources/the-new-stack.md>)

Topics: [Vercel](<https://devfeed.tech/topics/vercel.md>), [Secrets Management](<https://devfeed.tech/topics/secrets-management.md>), [Secure token management](<https://devfeed.tech/topics/secure-token-management.md>), [principle of ephemerality](<https://devfeed.tech/topics/principle-of-ephemerality.md>), [CI/CD secrets management](<https://devfeed.tech/topics/ci-cd-secrets-management.md>), [CI/CD](<https://devfeed.tech/topics/cicd.md>), [Deployment](<https://devfeed.tech/topics/deployment.md>), [credential-based attacks](<https://devfeed.tech/topics/credential-based-attacks.md>), [Single sign-on (SSO)](<https://devfeed.tech/topics/sso.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [agent-harnesses](<https://devfeed.tech/tags/agent-harnesses.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [api](<https://devfeed.tech/tags/api.md>), [api-keys](<https://devfeed.tech/tags/api-keys.md>), [api-management](<https://devfeed.tech/tags/api-management.md>), [application](<https://devfeed.tech/tags/application.md>), [applications](<https://devfeed.tech/tags/applications.md>), [attacks](<https://devfeed.tech/tags/attacks.md>), [authentication](<https://devfeed.tech/tags/authentication.md>), [authorization](<https://devfeed.tech/tags/authorization.md>), [ci-cd](<https://devfeed.tech/tags/ci-cd.md>), [credentials](<https://devfeed.tech/tags/credentials.md>), [deployment](<https://devfeed.tech/tags/deployment.md>), [developers](<https://devfeed.tech/tags/developers.md>), [ecosystem](<https://devfeed.tech/tags/ecosystem.md>), [marketplace](<https://devfeed.tech/tags/marketplace.md>), [security](<https://devfeed.tech/tags/security.md>), [static-credentials](<https://devfeed.tech/tags/static-credentials.md>), [token](<https://devfeed.tech/tags/token.md>), [vercel](<https://devfeed.tech/tags/vercel.md>)

### AI overview

Vercel is urging Marketplace vendors to replace persistent API keys with OIDC and short-lived, identity-based tokens, with "Not secure" listing warnings starting November 30 and a January 2027 compliance deadline. But its documented GitHub Actions deployment flow still requires a VERCEL_TOKEN secret; Vercel says OIDC support for its API and CLI is being extended. Its Connect service can issue scoped credentials at runtime and renew exchanged tokens, and now offers public-beta enterprise authorization via identity providers such as Okta. Connect still permits static keys for providers without identity-based authentication, so those underlying credentials may remain long-lived. The excerpt says Vercel did not publish the survey methodology, identify surveyed vendors, or say how many Marketplace partners must change.

### Source excerpt

Vercel wants to eliminate long-lived API keys from its ecosystem, and it is giving Marketplace vendors until January 2027 to The post Vercel moved to kill static credentials. Then developers spotted its own deployment token. appeared first on The New Stack.

## EMudhra Intermediate CA Certificates Compromised

DevFeed: [EMudhra Intermediate CA Certificates Compromised](<https://devfeed.tech/articles/emudhra-intermediate-ca-certificates-compromised-96032.md>)

Original publisher: [Read original article](<https://unit42.paloaltonetworks.com/emudhra-certificate-compromise/>)

Author: Unit 42

Published: 2026-10-22T17:15:32Z

Content type: article

Language: en

Sources: [Unit 42](<https://devfeed.tech/sources/unit-42.md>)

Topics: [certificates](<https://devfeed.tech/topics/certificates.md>), [Cryptography](<https://devfeed.tech/topics/cryptography.md>), [TLS (Transport Layer Security)](<https://devfeed.tech/topics/tls.md>), [interception](<https://devfeed.tech/topics/interception.md>), [network security](<https://devfeed.tech/topics/network-security.md>), [Secrets Management](<https://devfeed.tech/topics/secrets-management.md>), [Threat monitoring](<https://devfeed.tech/topics/threat-monitoring.md>), [Security](<https://devfeed.tech/topics/security.md>), [Certificate Transparency](<https://devfeed.tech/topics/certificate-transparency.md>), [principle of ephemerality](<https://devfeed.tech/topics/principle-of-ephemerality.md>)

Tags: [000](<https://devfeed.tech/tags/000.md>), [acme](<https://devfeed.tech/tags/acme.md>), [actions](<https://devfeed.tech/tags/actions.md>), [adversary](<https://devfeed.tech/tags/adversary.md>), [all](<https://devfeed.tech/tags/all.md>), [allows](<https://devfeed.tech/tags/allows.md>), [also](<https://devfeed.tech/tags/also.md>), [analysis](<https://devfeed.tech/tags/analysis.md>), [ca](<https://devfeed.tech/tags/ca.md>), [certificates](<https://devfeed.tech/tags/certificates.md>), [compromise](<https://devfeed.tech/tags/compromise.md>), [cryptographic](<https://devfeed.tech/tags/cryptographic.md>), [emudhra](<https://devfeed.tech/tags/emudhra.md>), [high-profile-threats](<https://devfeed.tech/tags/high-profile-threats.md>), [interception](<https://devfeed.tech/tags/interception.md>), [intermediate](<https://devfeed.tech/tags/intermediate.md>), [keys](<https://devfeed.tech/tags/keys.md>), [malware](<https://devfeed.tech/tags/malware.md>), [mitm](<https://devfeed.tech/tags/mitm.md>), [pki](<https://devfeed.tech/tags/pki.md>), [private](<https://devfeed.tech/tags/private.md>), [private-keys](<https://devfeed.tech/tags/private-keys.md>), [risk](<https://devfeed.tech/tags/risk.md>), [tls](<https://devfeed.tech/tags/tls.md>), [unit-42](<https://devfeed.tech/tags/unit-42.md>)

### AI overview

Eight eMudhra intermediate CA private keys were compromised after a certificate-generation defect produced insufficient randomness, potentially allowing attackers to derive the keys, issue unauthorized TLS certificates, and intercept or impersonate traffic. eMudhra revoked the eight certificates, and browser and OS vendors moved to revoke or distrust the intermediates. The guidance calls for identifying certificates and trust anchors tied to the affected CAs, reissuing affected certificates through another trusted CA, updating trust stores, and enforcing CRL and OCSP checks. It also recommends reviewing network, authentication, and identity logs for signs of interception or session replay, monitoring Certificate Transparency logs, restricting issuers with CAA records, and hardening automated certificate lifecycle management. Palo Alto Networks describes its Next-Gen Trust Security solution as supporting lifecycle management and emergency response; the text does not establish that it prevents this specific compromise.

### Source excerpt

Eight compromised eMudhra intermediate CA private keys risk unauthorized TLS signing and MitM interception. Read our analysis and containment guide. The post EMudhra Intermediate CA Certificates Compromised appeared first on Unit 42.

## Death to long-lived static credentials

DevFeed: [Death to long-lived static credentials](<https://devfeed.tech/articles/death-to-long-lived-static-credentials-96461.md>)

Original publisher: [Read original article](<https://vercel.com/blog/death-to-long-lived-static-credentials>)

Author: Malte Ubl

Published: 2026-10-22T04:00:00Z

Content type: opinion

Language: en

Sources: [Vercel News](<https://devfeed.tech/sources/vercel-news.md>)

Topics: [Secrets Management](<https://devfeed.tech/topics/secrets-management.md>), [Secure token management](<https://devfeed.tech/topics/secure-token-management.md>), [Vercel](<https://devfeed.tech/topics/vercel.md>), [Credential theft](<https://devfeed.tech/topics/credential-theft.md>), [principle of ephemerality](<https://devfeed.tech/topics/principle-of-ephemerality.md>), [secret rotation](<https://devfeed.tech/topics/secret-rotation.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [agents](<https://devfeed.tech/tags/agents.md>), [api](<https://devfeed.tech/tags/api.md>), [api-keys](<https://devfeed.tech/tags/api-keys.md>), [app](<https://devfeed.tech/tags/app.md>), [application](<https://devfeed.tech/tags/application.md>), [applications](<https://devfeed.tech/tags/applications.md>), [authentication](<https://devfeed.tech/tags/authentication.md>), [browser](<https://devfeed.tech/tags/browser.md>), [chat](<https://devfeed.tech/tags/chat.md>), [code](<https://devfeed.tech/tags/code.md>), [credential-theft](<https://devfeed.tech/tags/credential-theft.md>), [credentials](<https://devfeed.tech/tags/credentials.md>), [data](<https://devfeed.tech/tags/data.md>), [database](<https://devfeed.tech/tags/database.md>), [dependency](<https://devfeed.tech/tags/dependency.md>), [eff](<https://devfeed.tech/tags/eff.md>), [encryption](<https://devfeed.tech/tags/encryption.md>), [static](<https://devfeed.tech/tags/static.md>), [static-credentials](<https://devfeed.tech/tags/static-credentials.md>)

### AI overview

Long-lived API keys and database passwords can remain usable after leaks, and rotating them can disrupt production. The excerpt presents OIDC as an alternative: applications receive signed, short-lived identity tokens that services verify against trust policies, limiting the usefulness of a stolen token and enabling fresh credentials without distributing persistent keys. Vercel describes existing OIDC federation for temporary AWS, Google Cloud, and Azure credentials, plus runtime token access through Vercel Connect, which caches provider tokens and refreshes them near expiry. It also offers OIDC UI components for trust configuration and token verification. Token exchange can add network latency; Vercel says platform-issued tokens can be read from request headers and unexpired tokens reused. It reports that 8.4% of over 1,000 surveyed SaaS vendors supported secure authentication, and sets Marketplace deadlines: a "Not secure" listing label by November 30, 2026, and delisting in January 2027 for vendors lacking secure authentication.

### Source excerpt

Fifteen years ago, TLS was something you'd use to protect the most sensitive areas of your app. Then Google, Let's Encrypt, the EFF, and others led the push for HTTPS everywhere, making encryption the default for every request. Without it, anyone between the browser and server can read and tamper with the traffic. OAuth, OIDC, DBSC, and related standards already provide building blocks for secure authentication, but standards alone do not make secure methods the norm. HTTPS needed tooling that made encryption easy and a sustained push that made it expected. Machine-to-machine authentication still resembles HTTPS in 2010, with some vendors and enterprises leading the way while long-lived static credentials remain the default for many SaaS vendors and infrastructure providers. At Vercel, we want to make fast and secure the default across the web, including in our own services and the integrations in our Marketplace. We've built OIDC UI, a shadcn/ui registry that gives providers trust-policy forms and token verification code. We're expanding OIDC support across Vercel and setting new authentication requirements for Marketplace providers. Long-lived static credentials are insecure Long-lived static credentials are secrets such as API keys and database passwords that applications keep reusing, often with no expiration. As agents take on more useful work, we give them access to our systems, tools, and data through credentials that can grant permissions far beyond the task at hand. Those credentials can leak into a code repository or chat, be stolen when a service or dependency is compromised, or be exposed when prompt injection leads an agent to reveal them. However it happens, an attacker can use the credential with the permissions it grants until it expires or is revoked. Recent credential-theft and supply-chain incidents across the industry show why encryption at rest isn't enough when malicious code can read an application's credentials in production. Rotating exposed

## Best practices for integrating IAM Roles Anywhere with your PKI

DevFeed: [Best practices for integrating IAM Roles Anywhere with your PKI](<https://devfeed.tech/articles/best-practices-for-integrating-iam-roles-anywhere-with-your-pki-93081.md>)

Original publisher: [Read original article](<https://aws.amazon.com/blogs/security/best-practices-for-integrating-iam-roles-anywhere-with-your-pki/>)

Author: Abhishek Avinash Agawane

Published: 2026-10-21T17:03:50Z

Content type: article

Language: en

Sources: [AWS Security Blog](<https://devfeed.tech/sources/aws-security-blog.md>)

Topics: [AWS IAM](<https://devfeed.tech/topics/aws-iam.md>), [Amazon Web Services](<https://devfeed.tech/topics/aws.md>), [AWS Security Token Service](<https://devfeed.tech/topics/aws-security-token-service.md>), [identity and access management](<https://devfeed.tech/topics/identity-and-access-management.md>), [Secrets Management](<https://devfeed.tech/topics/secrets-management.md>), [Security](<https://devfeed.tech/topics/security.md>), [enterprise certificate authority certificates](<https://devfeed.tech/topics/enterprise-certificate-authority-certificates.md>), [Cryptography](<https://devfeed.tech/topics/cryptography.md>), [principle of ephemerality](<https://devfeed.tech/topics/principle-of-ephemerality.md>)

Tags: [access-management](<https://devfeed.tech/tags/access-management.md>), [advanced-300](<https://devfeed.tech/tags/advanced-300.md>), [amazon](<https://devfeed.tech/tags/amazon.md>), [amazon-web-services](<https://devfeed.tech/tags/amazon-web-services.md>), [amazon-web-services-aws](<https://devfeed.tech/tags/amazon-web-services-aws.md>), [aws](<https://devfeed.tech/tags/aws.md>), [aws-certificate-manager](<https://devfeed.tech/tags/aws-certificate-manager.md>), [aws-iam](<https://devfeed.tech/tags/aws-iam.md>), [aws-identity-and-access-management-iam](<https://devfeed.tech/tags/aws-identity-and-access-management-iam.md>), [best](<https://devfeed.tech/tags/best.md>), [best-practices](<https://devfeed.tech/tags/best-practices.md>), [credentials](<https://devfeed.tech/tags/credentials.md>), [getting](<https://devfeed.tech/tags/getting.md>), [iam](<https://devfeed.tech/tags/iam.md>), [iam-roles](<https://devfeed.tech/tags/iam-roles.md>), [identity](<https://devfeed.tech/tags/identity.md>), [identity-and-access](<https://devfeed.tech/tags/identity-and-access.md>), [identity-and-access-management](<https://devfeed.tech/tags/identity-and-access-management.md>), [pki](<https://devfeed.tech/tags/pki.md>), [practices](<https://devfeed.tech/tags/practices.md>), [roles](<https://devfeed.tech/tags/roles.md>), [security-blog](<https://devfeed.tech/tags/security-blog.md>), [security-identity-compliance](<https://devfeed.tech/tags/security-identity-compliance.md>), [with](<https://devfeed.tech/tags/with.md>)

### AI overview

IAM Roles Anywhere lets workloads outside AWS exchange CA-signed X.509 certificates for temporary IAM credentials. The article recommends distinct certificates and IAM roles per workload, trust policies that restrict access using certificate-derived subject, issuer or SAN attributes, and session policies to narrow permissions. Trust-anchor placement affects which certificates are accepted; short certificate validity limits exposure, while just-in-time issuance can apply role changes immediately but depends on CA support. Protect private keys with PKCS #11 stores or TPM 2.0 where available; macOS Keychain and Windows certificate stores are also supported. For credentials, SDKs refresh credential_process results in memory, but the AWS CLI invokes it on every command; `serve` avoids that repeated call, while `update` writes credentials to disk and carries file-access risk. The article also recommends expiry monitoring and blue/green trust-anchor rotation. Its sample deployment requires session duration not to exceed the IAM role's MaxSessionDuration and matching configured certificate attributes.

### Source excerpt

Using temporary credentials is a security best practice to reduce the risk from credentials getting compromised. With AWS IAM Roles Anywhere, now your workloads outside of Amazon Web Services (AWS) will be able to access AWS resources by obtaining temporary security credentials in AWS Identity and Access Management (IAM). IAM Roles Anywhere relies on public [...]

## AI Is Calling In Every Risk You've Accepted: Seven Insights From The AI Security Panel: Unfiltered

DevFeed: [AI Is Calling In Every Risk You've Accepted: Seven Insights From The AI Security Panel: Unfiltered](<https://devfeed.tech/articles/ai-is-calling-in-every-risk-you-ve-accepted-seven-insights-from-the-ai-security-panel-unfiltered-92994.md>)

Original publisher: [Read original article](<https://snyk.io/blog/ai-security-panel-unfiltered/>)

Author: Snyk Team

Published: 2026-10-21T00:00:00Z

Content type: article

Language: en

Sources: [Snyk](<https://devfeed.tech/sources/blog-rss-feed-snyk.md>)

Topics: [ai security](<https://devfeed.tech/topics/ai-security.md>), [Cybersecurity](<https://devfeed.tech/topics/cybersecurity.md>), [LLM security](<https://devfeed.tech/topics/llm-security.md>), [code security](<https://devfeed.tech/topics/code-security.md>), [alert triage](<https://devfeed.tech/topics/alert-triage.md>), [principle of ephemerality](<https://devfeed.tech/topics/principle-of-ephemerality.md>)

Tags: [accountability](<https://devfeed.tech/tags/accountability.md>), [agent](<https://devfeed.tech/tags/agent.md>), [agent-harnesses](<https://devfeed.tech/tags/agent-harnesses.md>), [agent-skills](<https://devfeed.tech/tags/agent-skills.md>), [agentic](<https://devfeed.tech/tags/agentic.md>), [agentic-ai](<https://devfeed.tech/tags/agentic-ai.md>), [agentic-ai-adoption](<https://devfeed.tech/tags/agentic-ai-adoption.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-adoption](<https://devfeed.tech/tags/ai-adoption.md>), [ai-security](<https://devfeed.tech/tags/ai-security.md>), [americas](<https://devfeed.tech/tags/americas.md>), [application-security](<https://devfeed.tech/tags/application-security.md>), [attacks](<https://devfeed.tech/tags/attacks.md>), [awareness](<https://devfeed.tech/tags/awareness.md>), [blog](<https://devfeed.tech/tags/blog.md>), [defense](<https://devfeed.tech/tags/defense.md>), [developer](<https://devfeed.tech/tags/developer.md>), [executive](<https://devfeed.tech/tags/executive.md>), [insights](<https://devfeed.tech/tags/insights.md>), [open-source-security](<https://devfeed.tech/tags/open-source-security.md>), [panel](<https://devfeed.tech/tags/panel.md>), [remediation](<https://devfeed.tech/tags/remediation.md>), [risk](<https://devfeed.tech/tags/risk.md>), [risks](<https://devfeed.tech/tags/risks.md>), [security](<https://devfeed.tech/tags/security.md>), [snyk-open-source](<https://devfeed.tech/tags/snyk-open-source.md>), [speed](<https://devfeed.tech/tags/speed.md>), [vulnerability-insights](<https://devfeed.tech/tags/vulnerability-insights.md>)

### AI overview

The panel's shared diagnosis is that AI makes attacks cheaper and faster, while the growing bottleneck is fixing vulnerabilities before attackers can chain them. Leaders described agents as a new attack surface because they act on their own and may accumulate access; they recommended giving each agent a distinct identity, job-scoped access that expires, and an audit trail. They argued for machine-speed defense: automate safe remediation and SOC actions, with people handling exceptions and judgment. The panel differed on how to supervise nondeterministic systems and who should set AI accountability rules. Suggested near-term steps include scanning organizational code and infrastructure with frontier models, auditing agent access, enabling strong multi-factor authentication, and improving the ratio of issues fixed to issues found. The article cites Snyk research finding flaws in 36.82% of 3,984 scanned agent skills, and reports that only about 10% of recorded vulnerability findings with known fixes were resolved.

### Source excerpt

AI is making attacks cheaper and faster, exposing risks organizations already carry. Four security leaders share why machine-speed defense and faster remediation are now essential.

## 5 signs that standing privileges are putting systems at risk

DevFeed: [5 signs that standing privileges are putting systems at risk](<https://devfeed.tech/articles/5-signs-that-standing-privileges-are-putting-systems-at-risk-90220.md>)

Original publisher: [Read original article](<https://1password.com/blog/signs-standing-privilege-risk>)

Author: info@1password.com (Gabriel Avner)

Published: 2026-10-19T00:00:00Z

Content type: article

Language: en

Sources: [1Password](<https://devfeed.tech/sources/blog-on-1password-blog.md>)

Topics: [identity and access management](<https://devfeed.tech/topics/identity-and-access-management.md>), [Secrets Management](<https://devfeed.tech/topics/secrets-management.md>), [principle of ephemerality](<https://devfeed.tech/topics/principle-of-ephemerality.md>), [LLM security](<https://devfeed.tech/topics/llm-security.md>), [Secret Scanning](<https://devfeed.tech/topics/secret-scanning.md>)

Tags: [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [credentials](<https://devfeed.tech/tags/credentials.md>), [kubernetes](<https://devfeed.tech/tags/kubernetes.md>), [saas](<https://devfeed.tech/tags/saas.md>), [secrets](<https://devfeed.tech/tags/secrets.md>), [security](<https://devfeed.tech/tags/security.md>), [tips-advice](<https://devfeed.tech/tags/tips-advice.md>), [unified-access](<https://devfeed.tech/tags/unified-access.md>)

### AI overview

The article identifies five signs of standing privilege risk: excess administrators, long-lived credentials, permissions broader than actual use, sensitive actions with too little friction, and limited visibility into access. It describes zero standing privileges as a way to replace permanent access with temporary, task-scoped, auditable elevation, and presents 1Password Privileged Access as a product for implementing these controls.

### Source excerpt

In most organizations, standing privileges don't show up all at once. They accumulate quietly. A role is added "temporarily." A contractor needs broad access to finish a project. A service account gets oversized permissions because no one has time to fine-tune them. None of these choices seem harmful in the moment, but over time they build into a privilege surface that's far too large to monitor and far too easy to misuse. The challenge is that standing privileges often look harmless until something finally goes wrong. The indicators are visible long before the incident, but only if you know where to look. Top indicators of standing privileges Here are five of the clearest signs that your environment is carrying more access risk than it should. 1. You have more admins than your environment actually needs Admin creep happens naturally. In fast-moving environments, the easiest way to unblock a task is to give someone full access. It works, it's quick, and that access is rarely revisited. Multiply that across cloud accounts, SaaS tools, Kubernetes clusters, and databases, and you end up with dozens or even hundreds of identities that hold far more privilege than they need. The problem isn't just security. Admin sprawl amplifies operational fragility. A single mistake or a single malicious action can ripple far beyond its intended scope; the blast radius from an incident expands because privileges were broader than they needed to be. If you can't articulate why every admin in your environment truly needs that level of access, that's a strong indicator that standing privileges have taken root. 2. Long-lived keys and credentials are still everywhere API keys, service account tokens, SSH keys, and automation secrets are some of the easiest paths into a cloud environment. They're also some of the least governed. Many businesses have: Credentials that never expire Keys embedded in pipelines or code Secrets used across multiple systems, including by AI agents Tokens with full

## Keep API Keys Out of the LLM: Credential Brokering for AI Agents

DevFeed: [Keep API Keys Out of the LLM: Credential Brokering for AI Agents](<https://devfeed.tech/articles/keep-api-keys-out-of-the-llm-credential-brokering-for-ai-agents-90679.md>)

Original publisher: [Read original article](<https://www.swytchcode.com/content/ai-agent-credential-brokering>)

Author: Swytchcode Team

Published: 2026-10-06T10:20:00Z

Content type: tutorial

Language: en

Sources: [Swytchcode](<https://devfeed.tech/sources/swytchcode.md>)

Topics: [Secrets Management](<https://devfeed.tech/topics/secrets-management.md>), [API keys](<https://devfeed.tech/topics/api-keys.md>), [Credential theft](<https://devfeed.tech/topics/credential-theft.md>), [LLM security](<https://devfeed.tech/topics/llm-security.md>), [principle of ephemerality](<https://devfeed.tech/topics/principle-of-ephemerality.md>), [token jacking](<https://devfeed.tech/topics/token-jacking.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agent](<https://devfeed.tech/tags/ai-agent.md>), [api](<https://devfeed.tech/tags/api.md>), [api-keys](<https://devfeed.tech/tags/api-keys.md>), [credentials](<https://devfeed.tech/tags/credentials.md>)

### AI overview

The article explains how API keys can enter an AI agent's context through prompts, tool parameters, errors, repository files, and logs, and how exposed credentials could be disclosed through prompt injection. It describes credential brokering, which attaches secrets to API requests outside the model after policy checks, and offers a checklist covering redaction, least privilege, key rotation, and separate development and production credentials. It also describes how Swytchcode handles credentials and notes that its broker does not replace a hosted flow for each end user's account authorization.

### Source excerpt

How API keys leak into an AI agent's context, five ways to handle credentials compared, and a checklist for brokering keys so the model only names the action and never sees the secret.

## Ephemeral testing

DevFeed: [Ephemeral testing](<https://devfeed.tech/articles/ephemeral-testing-64786.md>)

Original publisher: [Read original article](<https://lemire.me/blog/2026/10/05/ephemeral-testing/>)

Author: Daniel Lemire

Published: 2026-10-05T08:00:51Z

Content type: article

Language: en

Sources: [Daniel Lemire](<https://devfeed.tech/sources/daniel-lemire.md>)

Topics: [Integration testing](<https://devfeed.tech/topics/integration-testing.md>), [AI-assisted coding](<https://devfeed.tech/topics/ai-assisted-coding.md>), [Software Engineering](<https://devfeed.tech/topics/software-engineering.md>), [principle of ephemerality](<https://devfeed.tech/topics/principle-of-ephemerality.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [ai](<https://devfeed.tech/tags/ai.md>), [integration-testing](<https://devfeed.tech/tags/integration-testing.md>), [software](<https://devfeed.tech/tags/software.md>), [testing](<https://devfeed.tech/tags/testing.md>), [unit-testing](<https://devfeed.tech/tags/unit-testing.md>)

### AI overview

The article proposes ephemeral testing: use an AI agent to build and test a temporary application or additional layer on top of a software component. How well that software works can reveal issues in the component's API, invariants, defaults, errors, and documentation. The author describes repeating the approach with different agents and tasks to evaluate a foundation while prototyping possible features.

### Source excerpt

We have many ways to ensure software quality. Unit testing. Fuzz testing. Integration testing. And so forth. I'd like to propose a method that was unthinkable before: ephemeral testing. (Ephemeral is a fancy word for 'throw away' or 'temporary'.) You write your code. You build your software component. Or the AI agent does ... Continue reading Ephemeral testing

## Why MCP security is about permissions overhaul

DevFeed: [Why MCP security is about permissions overhaul](<https://devfeed.tech/articles/why-mcp-security-is-about-permissions-overhaul-109604.md>)

Original publisher: [Read original article](<https://www.webflow.com/blog/mcp-security>)

Author: Mohit Bansal

Published: 2026-09-09T00:00:00Z

Content type: article

Language: en

Sources: [Webflow Blog](<https://devfeed.tech/sources/webflow-blog.md>)

Topics: [Security](<https://devfeed.tech/topics/security.md>), [Access Control](<https://devfeed.tech/topics/access-control.md>), [MSP MCP](<https://devfeed.tech/topics/msp-mcp.md>), [credential-based attacks](<https://devfeed.tech/topics/credential-based-attacks.md>), [principle of ephemerality](<https://devfeed.tech/topics/principle-of-ephemerality.md>), [tenant data protection](<https://devfeed.tech/topics/tenant-data-protection.md>), [trust](<https://devfeed.tech/topics/trust.md>)

Tags: [access-controls](<https://devfeed.tech/tags/access-controls.md>), [accountability](<https://devfeed.tech/tags/accountability.md>), [agent](<https://devfeed.tech/tags/agent.md>), [agent-identity](<https://devfeed.tech/tags/agent-identity.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agent](<https://devfeed.tech/tags/ai-agent.md>), [authorization](<https://devfeed.tech/tags/authorization.md>), [confused-deputy-problem](<https://devfeed.tech/tags/confused-deputy-problem.md>), [credentials](<https://devfeed.tech/tags/credentials.md>), [mcp](<https://devfeed.tech/tags/mcp.md>), [mcp-security](<https://devfeed.tech/tags/mcp-security.md>), [permissions](<https://devfeed.tech/tags/permissions.md>), [redesign](<https://devfeed.tech/tags/redesign.md>), [security](<https://devfeed.tech/tags/security.md>)

### AI overview

The article argues that MCP security failures often stem from credentials with excessive or stale permissions, rather than protocol defects. It cites incidents involving GitHub repository access and Asana cross-tenant access, then recommends task-specific secrets, authorization for each action and user, and temporary credentials. Its review questions cover current credential scope, per-site or per-repository controls, whether agents inherit or bypass a person's permissions, attributable activity logs, and review steps before production changes. It also identifies an unresolved identity problem: agents often borrow human OAuth credentials, while standards for agent identity and task-linked token lifetimes remain early proposals. The article recommends treating agent lifespan as an input to access decisions, while noting that no clean model for balancing lifespan and access is established.

### Source excerpt

Most MCP breaches trace back to over-scoped credentials, not broken code. Here's what a real permissions redesign checks for.