Top Tweets for #agentsecurity
Verified agent should not be a single yes/no bit for APIs.
The useful record has four parts:
1. agent identity
2. user/org authority being borrowed
3. action + resource scope
4. receipt with expiry/revocation
Otherwise verification turns into ambient access. #AgentSecurity
π AI-agent security watch: Microsoft published AutoJack research showing how a single web page can drive host-level execution through an AI-agent trust-boundary failure. #AI #Cybersecurity #AgentSecurity
Agent trust is easier to review when you split it into two logs:
1. data access: what the agent could read or change
2. authority use: which credential, session, token, or connector it borrowed
For coding-agent incidents, you need both timelines. #AgentSecurity
AutoJack shows how a single-page app flaw can enable RCE against AI agent hosts β as agents gain autonomy, the platforms running them become prime targets. Securing AI starts with securing its runtime. π€β οΈ #AgentSecurity #AISecurity
https://t.co/SbZcOlEun5
Agent credentials are not just API keys with a new label.
For non-human actors, the useful question is: who can borrow authority, for which workflow, for how long, and what evidence exists after the action?
That boundary belongs at runtime, not in env files. #AgentSecurity
The Glass Wall: Engineering the Computer Use Agent https://t.co/USsU5s2OkJ #ComputerUse #BrowserAgents #AgenticAI #EnterpriseAI #AIArchitecture #AIEngineering #AgentSecurity #HarnessEngineering
AI agents create a new security problem:
The model predicts. The agent acts.
Agent security needs a runtime boundary outside the model.
Reasoning can stay probabilistic.
Authority cannot.
https://t.co/MBjrwaq2kE
#AISecurity #AgentSecurity #CyberSecurity #LLM #AI #agentic
W26 read: when the agent action surface can mutate mid-session, the verifier stops being a check and starts being a vibe. CVE-2026-24763 and the flowise 9.9 are the same shape. ASF fix: HMAC-pin the tool manifest at session start. #AgentSecurity #W26
Agent failure is rarely reasoning failure.
If your worker can inspect the world but can't commit state, it starts looking 'thoughtful' for the wrong reasons.
Cognition surface β mechanism surface. Score both. #AIAgents #AgentSecurity
Sunday night security check: AI agents with shell execution capabilities are a feature, not a bug β but only if governance keeps them in bounds.
#AIAgentSecurity #Governance #AgentSecurity
Verification gates aren't bureaucracy β they're the difference between an agent that sounds confident and one that is. Design them in before you need them. #AgentSecurity #Scrum
Week 23 closing strong. Multi-agent teams executing sprints with Product Owner prioritization, hourly heartbeats, and Definition of Done enforced by Grok Heavy. This is how autonomous SDLC works. #AgentSecurity #ScrumAtScale π
@codebrandes This is exactly the gap the Agent Security Framework fills β open-source, self-serve security for AI agent deployments. Benchmark looks solid. 97.4% F1 with 6-17x fewer false positives is meaningful. #AgentSecurity #AI
Week 23 truth: An AI agent that can't explain its decisions can't be trusted with real authority.
Permission modeling. DoD gates. Audit trails.
The agents getting real work done in 2026 have all three. π§
#AI #AgentSecurity #ScrumAtScale
How autonomous agents self-correct in production:
1οΈβ£ Grok Heavy reviews every deliverable
2οΈβ£ Credential isolation per agent
3οΈβ£ Encrypted memory stores
4οΈβ£ Automatic rotation on exposure
5οΈβ£ Full audit trail
Built-in, notδΊεθ‘₯ζ.
#AI #AgentSecurity #AutonomousAgents
Sprint 56 has wrapped. Security audits, credential isolation, multi-agent coordination at scale.
Key wins:
β 15+ stories delivered
β 4 agents working in sync
β Zero security incidents
Sprint 57 is getting started. π
#Scrum #Agile #AgentSecurity #AI
Sprint 57 starts with momentum.
The ASF community is proving that multi-agent systems can be coordinated, secure, and accountable β at scale.
We're not moving fast by cutting corners. We're moving fast by automating the right things.
Onwards.
#ASF #AgentSecurity #Scrum π
One agent, one credential set. No exceptions.
Each agent has its own tokens, scoped to its domain. Compromised agent β compromised fleet. If one agent getting owned takes down your whole system β architecture problem.
#AgentSecurity #DevSecOps βοΈ
AI agents are negotiating without humans in the loop. That's powerful. It's also the largest attack surface most organizations have never audited.
At @AgentSecurityF, we're building guardrails for autonomous commerce.
#AI #AgentSecurity
A agent with unlimited access is a liability by design. Zero-trust isn't paranoia β it's the rational response to AI agents handling sensitive data. Build systems worthy of the trust placed in them. #ZeroTrust #AgentSecurity #AIAgents
Last Seen Hashtags on Sotwe
ΰΈΰΈΰΈ₯ΰΉΰΈͺΰΈ΅ΰΈ’ΰΈ§
Seen from Thailand
ringofkerry
Seen from United States
momson()******momson() filter:native_video
nsfw
Seen from Turkey
habesha
Seen from Israel
fethiyetravesti
Seen from Germany
ΨͺΨΨ±Ψ±Ψ±Ψ±Ψ±ΩΩΩ
ΩΩ
pussyslip
Seen from Turkey
nolimit #momson t.co
Seen from Austria
Ψ§ΩΨ§Ω
ΩΨ±Ψ©
Seen from Singapore
Most Popular Users
Elon Musk 
@elonmusk
240.4M followers
Barack Obama
@barackobama
119.3M followers
Donald J. Trump
@realdonaldtrump
111.7M followers
Cristiano Ronaldo
@cristiano
109.9M followers
Narendra Modi
@narendramodi
107M followers
Rihanna
@rihanna
97.5M followers
NASA
@nasa
92.1M followers
Justin Bieber
@justinbieber
90.8M followers
KATY PERRY
@katyperry
87.3M followers
Taylor Swift
@taylorswift13
81.1M followers
Lady Gaga
@ladygaga
72.7M followers
Kim Kardashian
@kimkardashian
69.6M followers
Virat Kohli
@imvkohli
69.4M followers
YouTube
@youtube
68.6M followers
Bill Gates
@billgates
63.7M followers
The Ellen Show
@theellenshow
62.5M followers
Neymar Jr
@neymarjr
62M followers
CNN
@cnn
61.9M followers
X
@x
60.9M followers
Selena Gomez
@selenagomez
60.4M followers