Skip to content

Topic

AI Agent Governance

Frameworks and control planes for governing what autonomous agents are permitted to do at runtime.

Current stories

buildOne report1 publisher

Ruflo's 100-agent swarm for Claude Code starts in read-only mode with approval switched on

Cognitum One released a new Ruflo on October 13 that turns one coding assistant into a swarm of 100+ agents. The parts worth copying are the controls around the model: approval modes, cost visibility and data gating.

Publishers:dev.to

Reality

Evidence25
Adoption55
Hype gap+30
Incentives70
Confidence55
productOne report1 publisher

Stacklok's Mecatl trades the coding agent's unrestricted shell for a governed tool catalog

Stacklok open-sourced Mecatl, a coding-agent harness split into separate services for organizations that want to run hundreds of sessions on Kubernetes. The company argues that a harness built as one laptop process has to become a distributed system once agents run unattended.

Publishers:cncf.io

Reality

Evidence35
Adoption
Not scored
Hype gap+10
Incentives80
Confidence35
buildOne report1 publisher

AWS puts the agent runtime behind an n8n node, and leaves the permissions with you

An open-source AgentCore node moves memory, sandboxed execution and tool access into a managed harness. The install is the easy part; the access decisions do not move.

Publishers:dev.to

Reality

Evidence45
Adoption
Not scored
Hype gap+10
Incentives70
Confidence50
productOne report1 publisher

Salesforce's AIforce puts 37 prebuilt sales skills inside Claude

Salesforce is pushing its business logic and permissions into other companies' interfaces. For the admin rolling this out, the decision is which permission set a skill runs under when it fires from a chat window.

Publishers:siliconangle.com

Reality

Evidence33
Adoption21
Hype gap+44
Incentives84
Confidence41
buildOne report1 publisher

Flipping one deny to an allow broke the agent journal at record 3

The journal chains each record's hash into the next, so a one-line edit to a refusal surfaced as a numbered record; anyone able to rerun the writer can still recompute the chain and pass the check.

Publishers:dev.to

Reality

Evidence48
Adoption15
Hype gap+28
Incentives80
Confidence42
productOne report1 publisher

Salesforce exposes its platform to coding agents through more than 60 new MCP tools

Headless 360 makes capabilities that used to sit behind the Salesforce console callable as APIs, MCP tools and CLI commands, and more than 60 of the new tools point coding agents such as Claude Code and Cursor at live orgs.

Publishers:salesforce.com

Reality

Evidence28
Adoption26
Hype gap+42
Incentives92
Confidence55
productOne report1 publisher

Pigment's prompt-built forms inherit the row-level permissions of the model beneath them

Frames builds forms and scenario planners from a prompt and wires them into the live planning model, which means it inherits the access rights already sitting there along with any mistakes in them, and hands a human the job of certifying formulas an agent wrote.

Publishers:siliconangle.com

Reality

Evidence30
Adoption30
Hype gap+32
Incentives82
Confidence48
securityOne report1 publisher

Prompt injection recruits a fully credentialed agent

Pillar Security's Dor Sarig argues the non-human identity buildout answers who an agent is rather than what it does, and cites an internal agent any Slack message could trigger across a thousand private repositories.

Publishers:scworld.com

Reality

Evidence24
Adoption18
Hype gap+28
Incentives86
Confidence52
buildOne report1 publisher

Curator approval gates every record AWS Agent Registry shows a consumer

AWS now ships a first-party catalog for agents, tools, skills and MCP servers, moving the hard part off building them and onto whoever keeps the records current and clears the approval queue, a staffing question as much as a platform one.

Publishers:aws.amazon.com

Reality

Evidence46
Adoption28
Hype gap+24
Incentives88
Confidence57
productOne report1 publisher

Broadcom folds private AI into an integrated VMware Cloud Foundation stack

VMware Private AI Cloud stacks inference, deny-by-default agent sandboxes and token metering on Cloud Foundation 9, so the sovereignty pitch now arrives attached to an infrastructure contract platform teams already signed.

Publishers:siliconangle.com

Reality

Evidence32
Adoption28
Hype gap+38
Incentives84
Confidence52
buildOne report1 publisher

Running the agent server-side hands every ticket's commits to one shared service account

A dev.to writeup traces about sixty commits that landed under a colleague's git identity, and the server-side rebuild that followed fixed dependency drift while leaving the question of who authored what exactly where it was.

Publishers:dev.to

Reality

Evidence28
Adoption14
Hype gap+14
Incentives32
Confidence38
buildOne report1 publisher

Microsoft ships an MIT-licensed agent kernel: policy rings, Ed25519 identity, kill switch

Agent Governance Toolkit puts policy checks in the execution path rather than the prompt. The seam: the kernel is middleware inside the agent's process, so containers still do the isolating.

Publishers:dev.to

Reality

Evidence24
Adoption
Not scored
Hype gap+46
Incentives72
Confidence28