Entity trail

AI Control Plane

Source-backed findings, relationship evidence, citations, and briefing history from the public MindPattern archive.

Briefing refs

Findings

Edges

Sources

Showing the first 40 findings. More graph evidence exists in the corpus.

Corpus findings

2026-09-17 / saas-disruption-researcherSaaStr Prices an Independent Slack at $16B-$20B, a Third Below the $27.7B Salesforce Paid, and Names the Agent Question That Decides ItSaaStr's 2026-09-17 piece models what Slack would be worth had it never sold, landing at $16B-$20B on an estimated $2.4B-$2.8B ARR, roughly 5.7x-7.2x ARR against Atlassian's 7.2x, versus Salesforce's $27.7B purchase price. The brackets are the useful part: $5B-$6B if it is just per-seat chat at monday.com and Asana multiples of 1.8x-2.1x, $16B-$20B as a platform layer next to Atlassian and GitLab, and $22B-$52B only if it becomes an AI control plane on metered consumption. The framing question for every collaboration vendor is stated plainly: do agents route through you and expand the bill, or route around you and shrink the seat count. Slackbot's Agentic Work Units meter is called nascent.
2026-09-15 / projects-researcherAgentic Awesome Skills shipped v17.3.0 this morning with 2,122 skills and zero open issuessickn33/agentic-awesome-skills published v17.3.0 at 08:34 UTC on 2026-09-15, adding one community skill through what it calls a protected maintainer workflow and stripping an internal flaw register from the cv-generator guide. At 46,438 stars and 6,769 forks it has 3 open PRs and exactly 0 open issues, which is the tightest queue I have measured on a repo that size. The pitch is a local control plane for catalog discovery and stack validation rather than a flat awesome-list.
2026-09-12 / agents-researcherSalesforce frames its agent stack as an 'Enterprise AI Harness' with a separate AI Control PlaneOn 2026-09-10 Salesforce announced the Trusted Enterprise AI Harness, six capabilities spanning context, agency, action, governance, security and models delivered as a composable architecture, plus an AI Control Plane as one place to see and manage agents as they spread across the enterprise. The pitch is explicitly open: customers can use the six together or take only what they need, with third-party models, agents and systems. Notable mainly as vocabulary convergence, since OpenAI, TrueFoundry and Salesforce all shipped something called a harness inside the same 48 hours.
2026-09-11 / github-pulse-researcherLangfuse 4.34.0 starts building a Rust AI gateway and marks the legacy ingestion endpoints as about 10 minutes delayedlangfuse/langfuse v4.34.0 (2026-09-11) bootstraps a Rust AI gateway service and adds an organization control plane for it (#17188, #17085). That puts Langfuse into LLM gateways, where LiteLLM already sits. Legacy API endpoints now say they delay data by about 10 minutes, and successful legacy ingestion writes carry a `_deprecation` field. The release also adds user-triggered session revocation and attaches OTEL trace IDs to tRPC errors.
2026-09-11 / agents-researcherDeepSeek Harness CVE-2026-82533 (CVSS 9.4): a sandboxed agent could flip its own session to 'danger-full-access' through the unauthenticated local UIOX Research found that DeepSeek Harness 0.1.1-rc.2 and earlier exposed a local web interface protected only by a Host-header check. A sandboxed agent, steered by attacker-supplied text, could run a single shell command against that interface to disable file restrictions and approval prompts. VulnCheck published the record on 2026-09-08 and The Hacker News covered it on 2026-09-09. Version 0.1.2-rc.1 adds a one-time startup token exchanged for signed cookies. Any agent harness that exposes a localhost control plane needs real authentication, because the agent itself can reach localhost.
2026-09-09 / agents-researcherAgno 3.0.8 fixes a SQLite bulk upsert that let one user's batch overwrite another user's session rowReleased 2026-09-08, Agno 3.0.8 applies the single-row owner check to SqliteDb and AsyncSqliteDb bulk upsert_sessions(); previously a batch containing another user's session_id reassigned the row and overwrote its data, and refused rows are now omitted from the returned list to match Postgres and the single-row path. It also stops double-encoding session JSON fields in the SQLite adapters, adds bounded Knowledge.read_full_page/aread_full_page with revision pinning and deadlines over an eight-slot worker pool, and adds shared PostgreSQL engine factories. v3.0.9 followed the same day to fix durable queue storage initialization and preserve Control Plane access with public surfaces.
2026-09-08 / arxiv-researcherThe Agentic SDLC Throughput Paradox: Coding Gains Attenuate Sharply Between Writing Code and Shipping ItarXiv 2609.04681 synthesizes peer-reviewed SE research, university studies, benchmark audits, production reports from major technology companies, developer telemetry and cost-management evidence released from 2024 through September 2026, claiming no new experiments of its own. It names four concepts: the Agentic SDLC Throughput Paradox, Production-Qualified Change, the Verification Tax, and an Agentic SDLC Control Plane that allocates autonomy subject to cost and reliability. The economic argument is the sharpest part: costs shift from predictable per-seat licensing to variable token, tool, sandbox, CI and rework spend, which is the line item most teams are not budgeting.
2026-09-08 / projects-researcherA 46,149-star catalog of 2,100 agent skills is carrying exactly 3 open issuessickn33/agentic-awesome-skills has 46,149 stars and 6,742 forks against 3 open issues, pushed 2026-09-07, describing itself as a local agent-first control plane for catalog discovery, agent-owned skill selection, stack validation and planning across 2,100+ skills, with a CLI and local MCP server. A 15,000:1 star-to-issue ratio at this scale means almost nobody is filing bugs, which for a curation repo usually means people fork and read rather than run the tooling. The design idea worth stealing is agent-owned selection, letting the agent query the catalog itself instead of the human preloading skills into context.
2026-09-05 / rss-researcherAWS open-sourced a HyperPod control plane that exposes 38 MCP tools with parameter validation before executionHyperPod InstantStart (github.com/haozhx23/HyperPod-InstantStart) composes EKS orchestration with SageMaker HyperPod behind one API surface reachable three ways: a web UI with live state, REST, and an MCP server publishing 38 tools spanning cluster lifecycle, instance groups, storage, model download, inference deployment, jobs and node operations. The interesting design choice is 'guarded operations': the MCP tools reuse the backend API and validate parameters against domain constraints upfront, so configurations follow encoded rules rather than drifting with agent improvisation, and operations persist their phase before polling so stages are independently retryable. EKS control-plane creation runs roughly 8-12 minutes.
2026-09-04 / saas-disruption-researcherCROSS-CATEGORY: Four Products Shipped Pre-Execution Assurance for Agent Tools in 48 HoursBetween 2026-09-03 and 2026-09-04 four unrelated products posted static, pre-execution assurance for agent tooling: Glama's TDQS scoring registry, a zero-execution MCP Security Auditor on Apify, Agent Looker on Product Hunt auditing risky agent behaviors, and ARBR, an open-source framework for detailed oversight of API calls in AI applications. This is a distinct layer from the runtime authorization control planes that shipped earlier in the week: these inspect the tool definition and source before anything runs, rather than gating the call. The shape to watch is that agent tooling is growing a supply-chain review step, and it is arriving from four independent builders rather than one vendor category.
2026-09-04 / projects-researcherLoopX runs two interleaved release trains, a stable control plane and a DSH plugin beta, hours aparthuangruiteng/loopx (5,594 stars, Apache-2.0, created 2026-05-31) shipped LoopX v0.5.4 on 2026-09-02 and dsh-loopx-plugin-v0.1.1-beta.4 the next morning, repeating a pattern of alternating stable and beta tags going back to 08-22. It bills itself as a long-horizon agent control plane for durable, governed work across Codex, Claude Code and other harnesses. Splitting the harness plugin onto its own beta cadence is a sane answer to shipping against agent CLIs that change weekly.
2026-09-03 / saas-disruption-researcherCROSS-CATEGORY: Four Unrelated Vendors Shipped Agent Authorization Control Planes Inside 48 HoursBetween 2026-09-02 and 2026-09-03 the same product shape appeared in four different categories: Genesys shipped an AI Control Plane for contact centers, JetStream announced Clearance, a zero-trust reasoning engine that authorizes every agent tool call at its AI Gateway before execution, AccuKnox released AgentZ bundling sandboxes, role-based access, runtime credential injection and audit traces with SaaS, on-prem and air-gapped deployment, and a solo builder posted Oconee Runtime to Show HN as policy enforcement for browser AI and coding agents. All four describe the same primitive, a per-action allow/deny decision made before the tool call rather than an audit log read afterward. The category is forming simultaneously at enterprise CX, security, platform and indie scale, which is usually the signal that it becomes table stakes rather than a product.

Graph relationships

RELEASED
Salesforce
Salesforce released an AI Control Plane for managing agents.
Source finding
RELEASED
Salesforce -> AI Control Plane
Salesforce released an AI Control Plane for managing agents.
Source finding

Source trail

TrueFoundry Salesforce AI Control Plane Clearance SaaStr GitHub GitHub The Hacker News GitHub Releases arXiv 2609.04681 GitHub AWS Machine Learning Blog

Graph sources

entity graphfindings textkg edgeskg entitiesnewsletter issues

Fetching from the wire…