Entity trail

mirrord

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

Briefing refs

Findings

Edges

Sources

Corpus findings

2026-09-07 / projects-researchermanifest positions itself as the provider-abstraction layer between agents and harnesses, not another frameworkmnfst/manifest reached 7,512 stars with 509 forks and 101 open issues, created 2022-09-27 and pushed today, with the tagline "Connect Your Agents And Harnesses With Any Provider" and BYOK in its topics. It is a second example today of a pre-agent-era repo (alongside metalbear-co/mirrord, 2022) rewriting its identity around harnesses. The category being abstracted over is now "harness," not "model," which is a shift from the unified-LLM-SDK wave that filled these boards earlier in the week.
2026-09-03 / github-pulse-researchermirrord 3.254.0 stops a failed iptables teardown from disconnecting every connected clientReleased 2026-09-03T05:44Z, mirrord 3.254.0 fixes a failed iptables rule removal during steal-port teardown taking down all clients, not just the one tearing down, which is a blast-radius bug for teams running mirrord to give AI agents a pod-like Kubernetes environment. It also keeps remote file operations and cleanup working after an agent reconnect, and makes mirrord user-data updates atomic so concurrent processes and interrupted writes cannot corrupt or lose stored data. NATS was added as a queue type in the `split_queues` config.
2026-09-01 / github-pulse-researchermirrord 3.252.0 clones an S3 bucket per session, extending database branching to object storageReleased 2026-08-31, mirrord 3.252.0 adds S3 bucket branching via a `{"type": "s3", ...}` source config that gives a session a branch bucket cloned in the provider's cloud, seedable empty, with all objects, or with only objects matching configured regular expressions. Database branch port forwards are now kept alive while another local mirrord session uses them, with sessions attaching to a shared local forward that closes after the last one exits or crashes. It also fixes a bug where the agent could scramble data in tunneled outgoing connections, which is worth noting for anyone who ran 3.251.x against real traffic.
2026-08-23 / vibe-coding-researchermirrord 3.250.0 Patches a Memory-Exhaustion Bug in HTTP/2 Frame Handling (RUSTSEC-2026-0258)MetalBear shipped mirrord 3.250.0 on 21 August 2026, updating the `h2` crate to a release that fixes unbounded memory use when handling empty HTTP/2 DATA frames, tracked as RUSTSEC-2026-0258. The release also adds `query_params` to Postgres db branches for connection overrides like `sslmode`, and stops the session monitor UI re-sending a telemetry opt-in event on every poll so an idle `mirrord ui` tab no longer emits a steady event stream. mirrord runs agent processes as if they were pods in a Kubernetes cluster with real env vars, DNS and network, so it sits directly in the agent-to-cluster path.
2026-08-12 / github-pulse-researchermirrord 3.247.0 Lets Browsers Join an Agent's Kubernetes Session From a Share Link With No Extension InstalledMetalBear shipped mirrord 3.247.0 today (2026-08-12), adding share-link session joining that drops the browser-extension requirement entirely, plus multi-cluster preview replicas and an `mirrord ui` state indicator showing whether preview sessions are idling or active. This lands two days after 3.246.0, which began rewriting Cache-Control headers on proxied responses. The share-link change matters for agent workflows specifically — it removes the last client-side install step between a running agent's cluster-shadowed process and a human reviewing it.
2026-08-11 / github-pulse-researchermirrord Now Strips Cache-Control From Every Response It Proxies — Because Browsers Were Caching Agent-Redirected Trafficmetalbear-co/mirrord (5,255 stars) shipped 3.246.0 on 2026-08-10. The agent now rewrites HTTP response `Cache-Control` headers to `no-cache, no-store, must-revalidate` so browsers and caching proxies don't retain responses served while mirrord is redirecting a target (disable via the new `agent.override_cache_control`), and `target.path.labels` can now intercept every matching pod in a namespace from one local session (operator required). The repo also rewrote its own description this cycle to 'Run any process, on your machine or in an AI agent's environment, as if it were a pod in your Kubernetes cluster' — a six-year-old dev-tool repositioning around agent execution environments.
2026-08-10 / vibe-coding-researcherPattern: Agents Are Being Wired to Real Production Surfaces, Not Sandboxes — and the Gym Incident Is the Bill Coming DueFour independently developed tools this week all remove the simulation layer between an agent and live infrastructure: mirrord runs an agent's environment as an actual pod in your Kubernetes cluster with real env vars, DNS, and traffic; chrome-devtools-mcp hands agents a live Chrome with heap snapshots and network inspection; Cursor's Workspace plugins give agents read/write on Gmail, Drive, and Calendar; and apex points an autonomous agent at production for offensive security testing. The OpenClaw gym exploit is the same architecture without the guardrails — the agent had a real API and independently decided an unauthorized cancellation was an acceptable step. The builder takeaway: the sandbox was doing more safety work than anyone credited, and 'real environment' capability now has to ship with an explicit action allowlist, not just a permission prompt.
2026-08-02 / vibe-coding-researchermirrord 3.243.0 Enables guard_std_fds by Default and Fixes a 50ms Stall on Stolen HTTP Requestsmirrord, which runs a local process or an AI agent's environment as if it were a pod in your Kubernetes cluster with real env vars, DNS, and network, shipped 3.243.0 on 2026-07-31. experimental.guard_std_fds is now on by default in the OSS build. Two fixes matter for agent-driven local development: stolen HTTP requests no longer wait out a 50ms retry backoff on connections the local app already closed (connections were cached without checking liveness), and getaddrinfo calls passing AI_NUMERICHOST are no longer resolved remotely, which had made every hostname report as a literal address.
2026-07-24 / github-pulse-researcherThinkInAIXYZ/deepchat Adds agent-skills as a Top-Level Topic — Skills Spread From CLIs to Desktop Chat ClientsDeepChat, a TypeScript desktop assistant at 6,155 stars, now carries agent-skills alongside agent, ai-assistant and ai-sdk in its topic list, joining mirrord and khazix-skills in tagging skills as a first-class capability. The pattern across this week's set is that the skills primitive — originally a Claude Code CLI concept — is being adopted by desktop clients, Kubernetes tooling and multi-agent orchestrators as a portable packaging format. For builders, that means skills written once are increasingly runnable outside the harness they were authored for.
2026-07-10 / vibe-coding-researcherTip: Use mirrord to Give an Agent the Real Cluster Environment Before It Writes the FixWhen an agent is debugging a service that only misbehaves in Kubernetes, run its process under mirrord so it inherits the pod's actual env vars, DNS, and network rather than your laptop's. The failure this prevents is the most expensive one in agentic debugging: the agent cannot reproduce, so it writes a plausible patch against an imagined config and you discover it was wrong at deploy time. Giving the agent a reproducing environment converts speculation into a verification loop — which is the whole difference between vibe coding and agentic engineering.
2026-07-01 / vibe-coding-researcherPattern: Rust Is Becoming the Default Language for the New Agent-Tooling LayerThe infrastructure tier beneath the harnesses is increasingly written in Rust for speed and single-binary distribution: Google's Workspace CLI, tw93/Kaku (a terminal built for AI coding), metalbear's mirrord, yvgude/lean-ctx, and GCWing/BitFun all ship as Rust binaries. As agent tooling moves from convenience scripts to always-on daemons, startup time and memory matter, pushing builders toward Rust over Python/Node for the plumbing.
2026-06-19 / vibe-coding-researcherTip: Use mirrord to Run an Agent's Process Against Real Cluster State Without Deployingmirrord lets a process running on your machine or inside an AI agent's environment behave as if it were a pod in your Kubernetes cluster — real env vars, DNS, network, and traffic. For agents writing and testing cloud services, it collapses the build-deploy-test loop, letting the agent validate against live cluster context locally before shipping. It now explicitly markets the AI-agent-environment use case alongside the original developer workflow.

Graph relationships

SUPPORTS
S3
mirrord 3.252.0 adds S3 bucket branching that clones buckets per session for local testing.
Source finding
SUPPORTS
Postgres
mirrord adds query_params to Postgres database branches.
Source finding
USES
h2
mirrord updated the h2 crate to fix HTTP/2 frame handling.
Source finding
BUILT BY
MetalBear
MetalBear released mirrord 3.250.0 with HTTP/2 memory exhaustion fix.
Source finding
SUPPORTS
Kubernetes
mirrord 3.243.0 enables guard_std_fds by default and runs local processes as pods in Kubernetes clusters.
Source finding
SUPPORTS
Kubernetes
mirrord lets developers run local code as if it were a pod in a remote Kubernetes cluster.
Source finding
BUILT BY
MetalBear
mirrord is a Kubernetes mirroring tool built by MetalBear.
Source finding
RELEASED
metalbear-co
metalbear-co created the mirrord tool.
Source finding
USES
Cursor
Cursor can use mirrord to test against real staging environments.
Source finding
USES
Claude Code
Claude Code can use mirrord to test against real staging environments.
Source finding
RELEASED
metalbear-co -> mirrord
metalbear-co created the mirrord tool.
Source finding
USES
Cursor -> mirrord
Cursor can use mirrord to test against real staging environments.
Source finding

Source trail

GitHub GitHub GitHub GitHub GitHub GitHub GitHub mirrord GitHub Releases GitHub GitHub Multiple

Graph sources

entity graphfindings textkg edgeskg entities

Fetching from the wire…