Code & AI agent sandboxes: read the guide

AgentNest

Open source toolSandbox

The open-source runtime for secure AI agent execution. Disposable, policy-controlled environments for Python, shell, files, packages, browsers, GPUs, and Git work — self-hosted, Python-first, and deliberately not another cloud or cluster orchestrator.

Homepage Docs Repository

Scanned 23 Jul 2026 · Schema v5 · features v17

Licence: Apache-2.0
Self-hosted; you pay for the infrastructure.
Free tier: Free to run
Isolation: Container
shares the host kernel
Regions: —
Not disclosed on the public surface.
Capabilities: 11 / 54
1 partial · 20 fail · 7 not assessed
Agent readiness: 5/7 · 7/9
read · drive, pre-login surface

Advertised for

Disposable sandboxed execution of untrusted / agent generated Python and shell code: stateful Python REPL sessions, package installs (egress allowlisted), file work, Playwright browser automation (browser preset), GPU compute (gpu preset), and Git workspaces (clone/diff a repo, run tests). Used as a sandboxed code tool inside agent frameworks (LangChain/smolagents) or exposed over MCP.

Built for

Developers, then technical

· Python SDK + `agentnest` CLI; code-first README (`with Sandbox(...) as sandbox: ...`).
· Vocabulary: containers, isolation boundaries, gVisor/Kata/Firecracker, seccomp/AppArmor, capabilities, digests.
· Deep reference docs + runnable examples/; no GUI, no templates gallery.

01 Runtime Each fact carries its source; dots show confidence. Open a row for the evidence.

Isolation & lifecycle

Isolation primitive

Containershares the host kernel

Persistence

Ephemeral by default — one container + one host temp dir mounted at /workspace, both deleted on destroy / context exit / lifetime expiry / timeout. Read-only root + restricted tmpfs scratch. No arbitrary volume mounts (deliberate). Durable state is via `Sandbox.snapshot('workspace.tar')` / `restore` and `artifacts('output/**/*')` discovery, not a mounted persistent disk.

Snapshots / forks

Yes

Resources & limits

Storage

Ephemeral /workspace (host temp dir) + restricted tmpfs scratch; path-safe file I/O (rejects absolute paths and `..`; no-follow FDs defeat symlink escapes/races; 64 MiB read limit). No arbitrary volume-mount option by design. Durable output via snapshots (tar) + artifact discovery (globs with sha256).

Limits & quotas

Per-sandbox at creation: memory, cpus, PIDs, tmpfs scratch size, max_output_bytes, per-file read limit, and total lifetime + per-command timeout. Non-destructive timeouts (0.3.0): a slow command is killed on its own (in-container `timeout`) leaving the sandbox alive; a host-side backstop still destroys sandboxes on images without coreutils timeout. Bounded by the host you self-host on — no platform quotas.

GPU

Yes

Access

Deploy interface

Python SDK (embed `Sandbox`/`AsyncSandbox` in your process) + `agentnest` CLI (run/shell/doctor/backends/serve/mcp/demo/prune) + an MCP server (`agentnest mcp`) + an authenticated remote HTTP API server (`agentnest serve`, FastAPI) + YAML profiles (agentnest.yaml). Framework adapters: framework-neutral SandboxRunner, a LangChain StructuredTool, a smolagents executor.

SDK languages

Python only (Python-first, `>=3.10`). Sync `Sandbox` + `AsyncSandbox`. The package IS the SDK (`pip install agentnest`).

Network & secrets

Egress controls

Yes

Secrets

Explicit `Secret('value')` wrapper — redacted to [REDACTED] in captured stdout/stderr; injected as environment values. Only wrapped values are redacted (not transformed derivatives). Git credentials are never inherited from the host and must be injected explicitly, wrapped in Secret.

Bring your own cloud

Placement & use

Jurisdiction

Advertised workloads

Disposable sandboxed execution of untrusted / agent-generated Python and shell code: stateful Python REPL sessions, package installs (egress-allowlisted), file work, Playwright browser automation (browser preset), GPU compute (gpu preset), and Git workspaces (clone/diff a repo, run tests). Used as a sandboxed code tool inside agent frameworks (LangChain/smolagents) or exposed over MCP.

02 Capabilities 11 pass · 1 partial · 20 fail · 54 of 61 assessed

Supported — Documented on the public surface and verified.
Partially supported — Exists with caveats: gated, thin, or enterprise-only.
Not supported — Checked and not offered.
Not assessed — Scanned before this capability was tracked. Not a lack.

Compute & storage 3/6

Persistent diskStorage that survives restart/stop (not ephemeral-only).

Not supported

Object storageManaged S3-compatible / blob object store the platform PROVIDES to a user's workloads (buckets, presigned URLs). DISTINCT from persistent-disk (block/volume attached to an instance) and from a self-host dependency (source.self_host.deps[]=["object-storage"] is an infra dep to RUN an OSS tool — the opposite direction). SCOPE varies and matters: native first-class (Railway Buckets, Cloudflare R2) vs a sibling SKU / extension (Fly Tigris, Bunny Storage, DigitalOcean Spaces) — record which in the evidence quote / note. Admitted v15.

Not applicable

Snapshots / forkingSnapshot or copy-on-write fork of a running instance (filesystem-only counts). Whether the snapshot also restores memory + running processes is recorded separately by memory-snapshot (v19).

Supported

Live-state (memory) snapshotsA snapshot / checkpoint / pause captures a running instance's MEMORY and PROCESS state and restores it, so running processes continue where they left off instead of a freshly booted instance with the same files. Counts whether user-triggered (snapshot, checkpoint, pause-to-storage) or used to restore pre-initialised state (memory snapshots for fast cold starts). Distinct from snapshots-forks (any snapshot or fork, including filesystem-only; a product passing that key may still fail this one) and from scale-to-zero (an idle-cost property; suspension may or may not keep memory). Actor multiplexing (see scale-to-zero) records its live-state half here — only when the suspended actor's MEMORY/process state is what comes back, not a freshly booted worker with the actor's files.

Not assessed

Scale to zeroIdle instances suspend at no / negligible cost. ACTOR MULTIPLEXING (many suspended stateful actors sharing a small pool of pre-warmed workers, woken on request — Agent Substrate, AX, kagent; Rivet's hibernating actors) has NO key of its own (proposal REJECTED v20) — score the idle-cost half HERE and the live-state half under memory-snapshot; the oversubscription ratio / density is a performance characteristic the registry does not key.

Not applicable

GPU computeGPU-backed instances available.

Supported

Multi-regionCompute placeable in more than one geographic region.

Not applicable

Sandbox templates / custom imagesBuild reusable sandbox/VM images from an OCI image or Dockerfile (a base you fork instances from).

Supported

Cloud-init / ignition configCustomise a VM / microVM / instance at boot via cloud-init (or ignition) user-data — the same boot-metadata interface as AWS/Azure/GCP. Inject packages, users, SSH keys, files, and run-commands declaratively at launch. Distinct from sandbox-templates (a PREBUILT image you fork instances from) — this is PER-LAUNCH declarative customisation, not a baked image.

Not supported

Desktop / computer-use sandboxA virtual desktop (GUI) for LLM computer-use — browser + desktop automation, often via noVNC/VNC.

Not supported

Selectable isolation runtimeThe user chooses the isolation runtime per workload / environment / sandbox from more than one option (e.g. microVM vs gVisor vs plain container, native namespaces vs gVisor, gVisor vs VM). Records the CHOICE; the isolation_primitive facet still records the best-evidenced primary. A platform that switches runtime automatically without a user choice (e.g. gVisor forced for GPU) does not pass on that alone.

Not assessed

Networking & delivery 1/2

Automatic HTTPSTLS endpoints provisioned automatically per workload.

Not applicable

Built-in reverse proxyPlatform routes external traffic to workload services.

Not applicable

Custom domainsBring-your-own domain via CNAME/DNS.

Not applicable

Public preview URLsShareable per-deploy / per-branch URLs.

Not applicable

Direct public IPv6Per-instance stable public IPv6 for raw TCP/UDP (beyond a managed HTTPS proxy) — e.g. mosh / raw sockets.

Not applicable

Egress controlsConfigurable OUTBOUND network policy — egress firewall / allowlist / egress proxy on traffic LEAVING a workload. Scoped to OUTBOUND only (v15): private-network REACHABILITY between a user's own services is private-networking, not this. Static-egress-IP belongs here (an outbound concept).

Supported

Private networkingA platform-internal private network between a user's OWN services — service-to-service connectivity, VPC, service mesh, and/or service discovery, not exposed to the public internet. DISTINCT from egress-controls (OUTBOUND firewall/allowlist), reverse-proxy (PUBLIC inbound ingress), and direct-ipv6 (per-instance PUBLIC addressing). Admitted v15.

Not applicable

Email send / receiveSend and/or receive email from the workload.

Not supported

Edge KV storeA globally-replicated key-value store read from WITHIN an edge runtime at the PoP (low-latency, eventually-consistent) — Cloudflare Workers KV, Deno KV, Fastly KV Store, Akamai EdgeKV, Vercel Edge Config. Distinct from first-party-db (a managed DATABASE service you provision and connect to) and from http-query-api (HTTP access to an external store) — edge-kv is colocated with the edge compute and read at the edge.

Not applicable

Developer access 4/7

Public APIDocumented programmatic API.

Supported

OpenAPI specMachine-readable OpenAPI / Swagger spec published.

Not supported

CLIOfficial command-line tool.

Supported

SDKOfficial client library/libraries.

Supported

SSH accessDirect SSH into workloads.

Not supported

Filesystem APIRemote file operations over the API (list/read/write/watch/copy/upload) without SSH.

Supported

Git deployDeploy from a Git repo / push-to-deploy.

Not applicable

Private container registryHosted private image registry.

Not supported

E2B-compatible APIDrop-in compatibility with the E2B sandbox API/SDKs — existing E2B client code runs against this product by changing the endpoint/domain + key, with no business-code changes. E2B itself passes as the reference implementation. Distinct from public-api / sdk (the product has its own API — this key says it also speaks another vendor's) and from self-host-oss (a self-hosted E2B-compatible server is scored on both).

Not assessed

HTTP data APIQuery / read / write the stored DATA over plain HTTP/REST (or GraphQL) WITHOUT the native protocol driver (Postgres wire / Redis RESP) — agent- and edge-friendly data access via fetch(). The DATA plane; distinct from public-api (a control-plane / management API). Auto-generated-from-schema REST/GraphQL (PostgREST, pg_graphql) is a flavour of this.

Not applicable

AI & agents 1/9

Built-in coding agentShips a first-party AI coding agent in the product.

Not supported

Agent skillsInstallable / declarable agent skills or tools.

Not supported

AI app builderPrompt-to-app / vibe-coding generation surface.

Not supported

MCP serverExposes a Model Context Protocol server for agents.

Supported

MCP clientActs as an MCP client / consumes MCP tools.

Not supported

MCP-server hostingRun/host MCP servers as a workload on the platform (distinct from exposing a first-party MCP control server — that is mcp-server).

Not supported

LLM gatewayRoutes LLM API calls through the platform (multi-provider).

Not supported

BYOK for LLMsBring-your-own LLM provider keys.

Not supported

Model inference API (first-party)A FIRST-PARTY hosted LLM/model inference API — the provider serves its OWN catalogue of (usually open-weight) models behind a (typically OpenAI-compatible) endpoint, billed per token. Distinct from llm-gateway (a multi-PROVIDER proxy/router for YOUR calls to external providers like OpenAI/Anthropic) and from byok-llm (bring your own provider keys) — here the platform IS the inference provider, serving its own models.

Not applicable

Autonomous QA agentA first-party AUTONOMOUS AI agent that QA-TESTS a running app / implementation — exercising it like a real user (clicking through flows, checking behaviour) and reporting bugs/defects (often back to a coding agent, or before a human review). Distinct from built-in-agent (a code-AUTHORING agent — this one VERIFIES the result, it does not write the feature; a product can ship both), from ai-guided-code-review (organises a diff for a HUMAN to review — not an autonomous tester), and from observability (runtime metrics/logs, not active testing).

Not supported

Credential-injection proxyThe platform holds real secrets OUTSIDE the workload and injects them on the wire (an egress / broker proxy swaps a placeholder or phantom token for the real credential on matching outbound requests), so code running in the sandbox — e.g. an untrusted agent — never sees or can exfiltrate the real key. Distinct from secrets-management (storage + injection INTO the workload as env vars/files — the workload holds the secret) and from egress-controls (outbound allow/deny POLICY; a product can have an allowlist without injecting credentials, and vice versa).

Not assessed

Preinstalled coding agentThe platform installs and launches a THIRD-PARTY coding agent (a harnesses.yml entry — Claude Code, Codex, OpenCode, Pi ...) inside a workspace/sandbox it provisions, with the agent's auth wired in (your own account or API key), so you start an agent session rather than an empty box. Distinct from built-in-agent (a FIRST-PARTY agent the vendor ships) and from harnesses[] (which records that a harness CAN target the product, not that the product runs it for you).

Not assessed

Data 0/1

First-party databasePlatform-provided managed database.

Not supported

Database branchingCopy-on-write branch of a DATABASE (schema + data) as a dev-lifecycle primitive — branch your data like code, usually instant via a metadata operation. Distinct from snapshots-forks (forks a running compute INSTANCE / VM state) and from preview-environments (a full per-PR environment that may USE a DB branch). For a data product a CoW branch maps HERE, not to snapshots-forks.

Not applicable

Vector searchFirst-class storage, indexing, and similarity search of vector embeddings — a native vector type/index or a managed vector DB. The data-layer primitive for RAG / semantic retrieval.

Not applicable

Ephemeral instant DB (no-signup)Provision a working, usable instance via a single command / HTTP call with NO account, signup, or card — typically time-boxed (expires unless claimed). A zero-onboarding path an agent can hit directly.

Not applicable

Platform & operations 0/4

Scheduled tasksCron / scheduled job primitive.

Not supported

Durable workflowsDurable execution / long-running workflow primitive.

Not supported

Preview environmentsA full EPHEMERAL environment spun up per PR / branch — compute + addons + (often) a DB branch — torn down on merge. More than preview-urls (which is just a shareable URL for one deploy).

Not applicable

Built-in observabilityFirst-party metrics / logs / tracing / APM / profiling included in the platform (no third-party stack required to see what a workload is doing).

Partially supported

Managed CI/CD runnersPlatform-hosted, drop-in CI/CD runner VMs (GitHub Actions / GitLab), typically faster and cheaper than the CI vendor's own hosted runners, billed per runner-minute; isolated + ephemeral per job. Distinct from git-deploy (push-to-deploy of an APP) and scheduled-tasks (cron) — this is managed CI COMPUTE you point your existing pipelines at.

Not applicable

Project template marketplaceA marketplace / gallery of prebuilt PROJECT STARTER templates (apps or websites) a user clones or launches a new project from and then customises — the "start from a template" surface of a builder / PaaS. Distinct from sandbox-templates (a COMPUTE base IMAGE you fork runtime INSTANCES from — Dockerfile / OCI), from app-builder (the prompt-to-app GENERATION surface — a product can have both), and from one-click OSS-app catalogues / helm-addon-marketplace (deploy a FINISHED app as-is, not a starter you build on).

Not supported

Organisation & security 2/3

Built-in authPlatform-provided authentication / IAM for workloads.

Not supported

SSO / SAMLEnterprise single sign-on.

Not applicable

Secrets managementFirst-class secrets storage / injection.

Supported

TeamsMulti-user team accounts.

Not applicable

SharingShare a running workload/resource via link (optionally gated).

Not applicable

BYOCRuns inside the customer's own cloud account.

Not applicable

Deploy into customers' VPCsA SaaS vendor uses the platform to deploy and operate ITS software inside ITS CUSTOMERS' cloud accounts / VPCs (nested or "customer BYOC"), managed from the vendor's control plane. Distinct from byoc (your own workloads in your own cloud account).

Not assessed

Workload identityThe platform gives each workload its own verifiable identity (OIDC token, SPIFFE SVID, cloud IAM federation) that it exchanges for short-lived credentials, so no long-lived static secret is stored or injected. Distinct from secrets-management (stores / injects static secrets) and credential-injection-proxy (swaps real secrets in on the wire).

Not assessed

Self-host / OSS coreOpen-source core that can be self-hosted.

Supported

Proposed, not verified

Capabilities the scan saw but the registry does not track yet. A key is admitted only once two distinct products exhibit it. These are not scored.

Approval hooks (human-in-the-loop policy gate before execution/file-ops/snapshot/restore)approval-hooks
“Implement approve(ApprovalRequest) -> bool to gate execution, file operations, snapshots, and restore. Hooks run before the backend operation. Denial raises PolicyDeniedError.” github.com
Pluggable isolation backend (a RuntimeBackend contract; third-party backends via the agentnest.backends entry-point group — one API across Docker / gVisor / Kata / Kubernetes / remote Firecracker)pluggable-isolation-backend
“The public Sandbox API targets RuntimeBackend, not Docker. Publish a factory through the agentnest.backends entry-point group. Snapshot/streaming use runtime-checkable capability protocols; unsupported -> UnsupportedCapabilityError.” github.com

03 Agent readiness Pre-login surface · counts, not a grade

Can an agent read it?

parse-ability · docs, metadata, machine-readable surfaces

5/7

4 full · 1 partial · 71% of applicable checks

llms.txt
Not supported
“No llms.txt; the advertised docs site is not deployed (404).”
llms-full.txt
Not supported
“No llms-full.txt.”
robots.txt AI policy
Not applicable
“No live own web property — surface is github.com + pypi.org (their robots, not the product's).”
JSON-LD
Not applicable
“No own site to carry SoftwareApplication/Product json-ld (docs site not deployed).”
Sitemap
Not applicable
“No live own domain; the mkdocs sitemap would exist only if the docs site were deployed (it 404s).”
Meta tags
Partially supported
“GitHub repo OG card + PyPI metadata carry a clear title/description, but no product-site <head> to grade.”
Markdown negotiation
Supported
“All docs are in-repo markdown served clean via raw.githubusercontent.com (docs/*.md); the README itself is markdown.”
Content efficiency
Supported
“GitHub-rendered README + raw markdown docs are near-pure text (very high readable-text ratio).”
Docs discoverable
Supported
“README links quickstart/architecture/deployment/security/integration guides + examples/; comprehensive in-repo docs, first-class.”
Machine-readable pricing
Not applicable
“No hosted pricing — self-host is free under Apache-2.0.”
No anti-bot wall
Supported
“No challenge on public content; curl/gh fetched repo + raw docs cleanly.”

Can an agent drive it?

act-ability · APIs, SDKs, MCP, credentials, webhooks

7/9

3 full · 4 partial · 78% of applicable checks

Public API
Supported
“Documented Python API (mkdocstrings api.md) + a versioned authenticated remote HTTP API (`agentnest serve`).”
OpenAPI spec
Not supported
“No published/fetchable OpenAPI spec (FastAPI auto-serves one at runtime, not documented/shipped).”
MCP server
Supported
“pip-installable MCP server with docs: `agentnest[mcp]` + `agentnest mcp` + client config examples — earns full per rubric v3 (installable MCP server with docs).”
SDK packages
Partially supported
“One official SDK (Python only).”
Auth documented
Partially supported
“Remote API uses a bearer token (AGENTNEST_API_TOKEN); deployment guide says authenticate/terminate TLS/rotate tokens — a basic key-auth mention, no OAuth/scoped-key model.”
Self-serve credentials
Not applicable
“Self-host — no hosted signup/credential issuance (you set your own API token).”
Webhooks
Not supported
“No HTTP webhooks (audit events are in-process structured events, not signed webhook callbacks).”
CLI tool
Supported
“Official `agentnest` CLI.”
Error model
Partially supported
“Documented stable exception hierarchy (exceptions.py — ExecutionTimeoutError/PolicyDeniedError/UnsupportedCapabilityError), but not a documented JSON error model for the HTTP API.”
Rate limits documented
Not applicable
“No hosted API to rate-limit; per-sandbox resource limits are self-configured, not platform limits.”
Agent config files
Partially supported
“MCP server + framework adapters (LangChain/smolagents) + an 'AI agents' docs guide — one/two integration surfaces, but no agents.md / .well-known/agent / skills listing.”

04 Coding harnesses Which agents and IDEs are documented to run here

Advertised integrations 2

Claude Code Anthropic
Supported
“The MCP server exposes sandbox lifecycle and execution as tools any MCP client (Claude Code, Cursor, Claude Desktop) can call: `pip install 'agentnest[mcp]'; agentnest mcp`. Register in a Claude Code / Cursor MCP config: { mcpServers: { agentnest: { command: agentnest, args: [mcp] } } }.”
Source github.com
Cursor Anysphere
Supported
“any MCP client (Claude Code, Cursor, Claude Desktop) can call ... Register it with a client — for example in a Claude Code / Cursor MCP config.”
Source github.com

Plausible compatibility 0

Derived, not stored: the product passes SSH access, so any CLI coding agent is installable inside it. This is plausible compatibility, not a supported integration.

No general shell access documented, so nothing is derived. Absence is not unsupported.

05 Compliance & regions Standards held, with ordinal rigor; never a score

Certifications held 0

No attestations found on the public surface. Empty means nothing is advertised, after checking, not that nothing was checked.

Regions —

No region list captured.

06 Open source Apache-2.0 · fetched by machine, not quoted from marketing

Licence & repository

Licence: Apache-2.0OSI
Primary language: Python
Stars: 8
Forks: 0
Created: 17 Jul 2026
Archived: No

Activity

Last commit: 21 Jul 2026
Latest release: 0.3.0 · 21 Jul 2026
Issues open / closed: 0 / 0

No GitHub Releases or git tags (releases totalCount 0). PyPI is the release channel: 0.2.0 (2026-07-17) → 0.3.0 (2026-07-21), i.e. rapid early iteration over a single week. Development Status :: 3 - Alpha; SemVer promised after the 0.x series. CHANGELOG documents 0.1.0/0.2.0/0.3.0.

Fetched via gh on 23 Jul 2026.

Self-hosting

Methods: package-managerdockerkubernetessource-build
You operate: —

Free under Apache-2.0 + your own infra. `pip install agentnest` (package-manager); optional extras `agentnest[kubernetes|server|mcp|observability|all]`. REQUIRES a container substrate you provide: a Docker daemon (default backend, `docker>=7.1` is the only hard dependency) OR a Kubernetes cluster (`[kubernetes]` extra). No external service deps (no DB/Redis/object-store). GPU sandboxes need the NVIDIA Container Toolkit on the host; stronger isolation needs a host-installed gVisor (runsc) or Kata runtime; the Firecracker backend needs a remote microVM worker on Linux/KVM. source-build via `pip install -e '.[dev,docs]'`. Docker-compose.yml + Dockerfile.dev are for the dev/test loop and mount the Docker socket (trusted machine only).

Documentation signals 6/8

SupportedQuickstart
SupportedAPI reference
SupportedSelf-host guide
SupportedChangelog
Not supportedOpenAPI
Not supportedllms.txt
SupportedExamples
SupportedArchitecture doc

07 Company & presence Facts only, not scored

Company

Headquarters: Not disclosed — independent OSS project, no company entity.
Founded: Repo created 2026-07-17; first PyPI release (0.2.0) same day.
Team size: Single maintainer (github.com/mihirahuja1); authors listed as 'AgentNest contributors'. Small ~4k-LOC codebase ('auditable in an afternoon').
Hiring: No — independent OSS project, no careers page.

Activity

Last release: PyPI 0.3.0 on 2026-07-21 (no GitHub releases/tags). Prior 0.2.0 on 2026-07-17. pypi.org
Last blog post: No blog. github.com

Presence 2

Links found on the public surface scanned; not an exhaustive search.

08 Sources & method How this entry was produced

Public surface only

Everything here was read from github.com/mihirahuja1/agentnestOSS's unauthenticated pages, docs, pricing and repository on 23 Jul 2026. Nothing behind a login was used.

Statuses, not scores

Capabilities are pass, partial or fail against a written rubric (features v17). “Not assessed” means the scan predates the key.

Confidence is separate

Each fact carries its own confidence. Low-confidence facts are shown, never hidden, so you can weigh them.

Schema: scan v5
Features registry: v17
Keys tracked today: 61
Assessed on this entry: 54

Something wrong on this page?

Nearby in the atlas

All Sandbox products →