Repository navigation
Releases
Cloudseed v0.2.3
Cloudseed v0.2.3
Cloudseed now records usage for its own agent tasks and MCP calls, with the same report available through the CLI, agents, MCP and web console. Agent report review also works when cs or cloudseed was not originally on PATH.
What changed
- Cloudseed-only usage: run
cs usage, call MCPcloudseed_usage, or open Agents & MCP → Cloudseed usage. Reports show task status, available token counts, model provenance and explicit reasons for missing measurements. Filters, pagination and JSON export are available. - Optional ccusage analysis:
cs usage installinstalls the pinned, verified ccusage 20.0.24 native analyzer.cs usage report --engine ccusageestimates cost from Cloudseed-owned metadata and offline prices. It does not import unrelated conversations or search global agent history. - Simpler agent execution: the separate compression proxy, its runtime dependencies and its controls have been removed. Cloudseed uses the agent's normal transport. The optional Context brief is Cloudseed's task context summary.
- Reliable command access: private
csandcloudseedlaunchers point to the exact running installation, with an absolute path fallback. - Complete saved-report review:
cs evidence list/readand MCPcloudseed_evidenceprovide redacted reports and diagnostic logs through revision-bound pages. Agents can read every finding, observation timestamp, coverage limit and next action without starting another scan. - More useful scan diagnostics: new cloud scans retain bounded, redacted diagnostic excerpts with omission counts. Zero-error counters and ordinary prose mentioning errors no longer create false incomplete verdicts.
- More resilient Mac signing: bounded retries handle recognized transient Apple timestamp-service failures; publisher, hardened runtime and secure timestamp checks remain required.
Start using it
After installing v0.2.3, restart the Cloudseed console and MCP server so they use the new executable. New Cloudseed-launched agent tasks populate the usage report:
cs usage
cs usage install
cs usage report --engine ccusageThe native report works without installing the optional analyzer. Existing sessions cannot retroactively supply missing measurements. MCP reports calls, duration and response sizes; model tokens remain unavailable because the MCP client does not send them to Cloudseed. Cost estimates are neither invoices nor subscription-quota measurements.
Read the usage guide, agentic guide, MCP report workflow, and updated scenario 14.
Validation
- Release workflow: all 10 jobs passed at commit
14adc694c2d2c75a2e5dbb12f53fbcda707b4fe9. - Source suite: 3,960 tests in each of six Linux/macOS and Python combinations, with explicit optional/platform skips. All 19 offline/local scenarios passed (725 checks; 62 live/credentialed skips). Terraform: seven roots validated, 124 mocked cases passed.
- Every released binary passed 25 runtime tests plus five real native ccusage tests; both containers passed 23 plus five. Release-platform scenario checks totaled 1,500, with 60 explicit skipped-action occurrences.
- Independently downloaded all 22 assets; verified checksums, manifests, inventories, all 16 file attestations and both container attestations against the exact source/workflow/tag.
- Both downloaded Mac binaries passed strict Developer ID signature checks for Nimesh Pandeya, team
QPF2VF2885, hardened runtime and secure timestamps. - The downloaded Apple Silicon binary passed seven targeted usage/report tests and a real Claude client exercise using a synthetic local API. Complete report paging, redaction and exact recorded token counts passed. Its own release verifier also returned PASS.
- Updated GitHub Pages deployed from the same source; seven live endpoints and 54 content checks passed.
Automated source, package and offline scenario checks do not establish live deployment or recovery acceptance across all clouds. Developer ID signing is separate from Apple notarization; this release does not claim notarization.
Cloudseed v0.2.1
Cloudseed v0.2.1 fixes first-time public-IP detection, Cloudseed installation on bastions and private-cluster tunnel setup. It also adds explicit host-kubeconfig administration and clearer security scan evidence across the CLI, MCP, agents and web console.
Bastion and cluster access
Standalone releases now include portable Python source and the launcher needed to install the same Cloudseed version on Linux bastions, including from a macOS controller. Normal provisioning installs and verifies both cloudseed and cs, makes them available in login/SSH shells, and leaves unrelated existing commands untouched. Kubernetes-enabled bastions receive checksum-verified kubectl, Helm and k9s when tool provisioning is enabled, including VMware. Repair an older bastion from the original controller with:
cs provision CLOUD --env NAME --host bastionReplace CLOUD and NAME with the existing environment. Installation does not copy the controller's environment records, Terraform state, private SSH keys, kubeconfig or cloud credentials. Keep infrastructure lifecycle and managed platform workflows on the original controller; installing the CLI does not grant cloud or cluster permissions.
After explicitly configuring an authorized kubeconfig on the bastion, use cs kubectl --local-context get nodes, cs helm --local-context list -A, or terminal cs k9s --local-context. This mode validates the selected context and an HTTPS endpoint with TLS verification, and rejects Cloudseed cloud/environment selectors. kubectl/Helm also offer local_context=true through MCP and an explicit host-kubeconfig option in the console; those paths and built-in agent execution require target confirmation even for reads. The target is the machine running that interface. The mode does not fetch credentials, open a Cloudseed tunnel, import infrastructure state or create Cloudseed Undo entries.
On the original controller, first use of a private-cluster tunnel now prepares kubectl before updating the environment's kubeconfig, including commands launched through Helm or k9s. A failed kubeconfig update stops the command with recovery guidance instead of reporting the tunnel ready. TLS verification against the original API name stays enabled.
Fixed
- Bundled trusted CA roots for verified HTTPS during AWS/GCP/Azure public-IP discovery and downloads. Explicit certificate overrides remain supported; TLS verification stays enabled.
- Cloud scans retain every failed/manual observation, including lower severities, actual explanations, check IDs, affected resources, remediation and account/project/subscription scope.
- CIS/STIG, Kubernetes posture, image, host and FIPS scans preserve evidence and report empty, partial, malformed or unverified results as incomplete. Kubernetes-disabled architecture assessments no longer demand cluster diagnostics.
- Informational and explicitly excluded/inapplicable checks do not create artificial incomplete results. Unknown findings explain the specific evidence gap and how to resolve it.
- Detailed web report cards, evidence and remediation, explicit preview limits, full JSON/Markdown views and narrow-screen wrapping.
- macOS packaging uses Apple Developer ID signing for Nimesh Pandeya, team QPF2VF2885, and verifies the expected publisher, hardened runtime and secure timestamp before testing and attesting the signed bytes. Signing is separate from Apple notarization, which this workflow does not perform.
- Release provenance verification no longer combines mutually exclusive GitHub CLI flags.
Result policy
FAIL exits 1; INCOMPLETE exits 3; PASS and N/A exit 0. A failed result takes precedence over incomplete evidence. Cloud/CIS/host scans fail on any failed check. Kubernetes posture retains its High/Critical failure threshold; image scans retain their Critical threshold. Lower-severity findings remain visible, and every report explains the applicable policy and coverage limits.
scan architecture --max-age-days 30 accepts saved evidence up to 30 days old. Run scan cloud first, then architecture to consume that evidence. Neither a saved configuration check nor a benchmark pass establishes complete framework compliance. Existing reports are preserved; rerun scans to obtain the expanded report format.
Documentation
- Bastion installation and repair
- Kubernetes access and bastion authorization
- Scan walkthrough, including base AWS
- Well-Architected assessments
- Web console report guide
- Public-IP troubleshooting
Downloads and verification
Choose the executable for your OS and CPU: Linux/macOS, amd64/arm64. Each executable has checksums, a manifest and a complete SPDX software inventory. GitHub release-workflow attestations authenticate the build and artifact bytes.
The container image is ghcr.io/nimeshbuilds/cloudseed:v0.2.1, with amd64 and arm64 variants. Use the versioned tag; this release does not publish a latest tag.
After replacing a local binary, restart the Cloudseed console/MCP service so it loads the updated code and web assets. See the release verification scenario before running downloaded artifacts.
Validation
The release is built from fb546c119fd5be75e4461c7e084a2cc41723db55. All 17 merged-main checks and all 10 release jobs passed. The source suite ran 3,840 tests in each of six Linux/macOS and Python-version combinations. All 19 offline scenarios passed with 712 checks; 61 live or credential-dependent steps were explicitly skipped. Each of the four published binary platforms passed 23 packaged tests, and both container architectures passed 21 packaged tests, plus their release scenario checks.
Post-publication verification passed for all 22 downloaded release assets, all 16 file provenance subjects, both image attestations and the multi-architecture container index. Both downloaded Mac binaries passed strict Developer ID signature verification. The downloaded Mac ARM64 executable passed real HTTPS public-IP discovery for AWS/GCP/Azure, nine isolated native kubectl context/TLS checks, and Cloudseed's own checksum-and-attestation verifier.
These checks did not create, destroy or reprovision cloud resources. Automated tests and installation smoke checks do not establish live deployment acceptance across all clouds; live AWS/GCP/Azure/VMware deployment, workload and recovery validation still requires authorized environments and identities.
Cloudseed v0.2.0
Cloudseed v0.2.0 brings the changes since v0.1.0 into versioned downloads: Well-Architected assessments, shared operational workflows, runtime reliability fixes, and expanded documentation.
Highlights
- Well-Architected assessments:
cs scan architectureevaluates saved configuration and local evidence for AWS, GCP, and Azure, plus common architectural guidance for VMware. The CLI, console, MCP, and agent skills expose the same findings, remediation, and evidence limits. - Operational readiness: health and network diagnostics, deployment profiles, portable specifications, budget and plan guards, explicit expiry cleanup, drift checks, guarded upgrades, selected-application recovery, and an opt-in cloud acceptance harness.
- Runtime reliability: fixes for detached bundled services, VMware recovery state, environment locking, SSH/OpenVPN process detection, MCP cleanup, and bounded console output.
- Interfaces and documentation: 48 MCP tools, 19 step-by-step scenarios with CLI/agent/MCP/UI routes, refreshed reference pages, and the white-and-blue documentation design.
- Release verification: checksums, manifests, complete SPDX dependency inventories, and GitHub build-provenance attestations for the tested artifacts.
Downloads
Choose the executable for your operating system and architecture:
| Platform | Executable |
|---|---|
| macOS Apple silicon | cloudseed-darwin-arm64 |
| macOS Intel | cloudseed-darwin-amd64 |
| Linux ARM64 | cloudseed-linux-arm64 |
| Linux x86-64 | cloudseed-linux-amd64 |
The executable bundles Cloudseed and Terraform. Each has a matching .SHA256SUMS, .manifest.json, and .sbom.spdx.json file. Verify the downloaded executable and its GitHub provenance before running it, following scenario 19.
The container includes the cloud and Kubernetes tools and supports Linux amd64 and arm64:
docker pull ghcr.io/nimeshbuilds/cloudseed:v0.2.0
docker run --rm ghcr.io/nimeshbuilds/cloudseed:v0.2.0 --versionContainer inventory manifests bind each SPDX inventory to its architecture-specific immutable image digest. The matching container-amd64.* and container-arm64.* files are included below.
Compatibility and verification scope
Source installations support Python 3.9 and later. Packaged binaries and containers use Python 3.14; building the VMware provider requires Go 1.25 or later. The VMware Terraform provider has an independent version.
Release checks cover source tests, Terraform validation and mocked Terraform tests, dry-run scenarios, and native Linux/macOS binaries and Linux containers. These checks do not establish real AWS/GCP/Azure deployment acceptance. Architecture assessments inspect local configuration and saved evidence; missing or manual-review evidence remains explicit.
cloudseed v0.1.0
The first public release.
Added
- Landing zones on three clouds, one command each.
cloudseed setup aws|gcp|azurebuilds a VPC/VNet with public,
private (and on AWS isolated data) subnets, NAT, flow logs, a hardened bastion reachable only from your IP, and a
security baseline: CloudTrail, GuardDuty, Access Analyzer, EBS default encryption and optional Security Hub on AWS;
log retention and Data Access audit logs on GCP; Activity Log to Log Analytics and optional Defender on Azure.
Remote state in versioned, encrypted, private S3 / GCS / Azure Storage, or local state. - The same shape on your laptop.
cloudseed setup vmwarecreates a NAT'd bastion and private VMs on VMware
Fusion Pro 13+ or Workstation Pro 17+ from official Ubuntu or Debian cloud images, through cloudseed's own Terraform
provider (providers/vmdesktop). - Kubernetes. Private EKS, GKE and AKS with workload identity and control-plane logging, or RKE2 / kubeadm on
VMware;cs node add|list|remove|scale;cs k8s kubeconfig;cs kubectl,cs helmandcs k9swith a per-environment
kubeconfig and an automatic SSH tunnel through the bastion for private endpoints;cs env useto pick the cluster. - Platform catalog.
cs platform installwith 73 pinned items in 10 groups (basek8s, scaling, data, ai, agentic,
finops, devsecops, security, resilience, chaos): ArgoCD, kube-prometheus-stack, Loki + Alloy, OpenTelemetry,
cert-manager, Gateway API + Envoy Gateway, external-secrets, KEDA, Karpenter, MinIO, CloudNativePG, Strimzi, Trino,
KServe, vLLM, Ollama, kagent, Istio, Falco, Velero, Chaos Mesh and more. Cloud prerequisites (buckets, identities)
are created by the environment's own Terraform stack;cs platform uiexposes dashboards;cs platform template gitlab-ciwrites a CI pipeline. - Resilience.
cs dr backup|restore|schedule|testwith Velero, including an automated restore drill that verifies
every object and a volume's data and reports the measured restore time. - Chaos engineering.
cs chaos run basic|network|stress|fullwith Chaos Mesh, steady-state hypotheses and
PASS / FAIL / INCONCLUSIVE verdicts, against a canary or your own Deployment (--target), with JSON and Markdown
reports. - Compliance and security scans.
cs scan cis|kube|images|host|stig|cloud|fips|allwith kube-bench, kubescape,
trivy, OpenSCAP (CIS and DISA STIG profiles) and prowler, with saved reports. - FIPS 140 mode (
--var fips_mode=true): FIPS endpoints, FIPS node images, FIPS-only SSH and VPN algorithms,
RSA-4096 keys, TLS 1.2+ with FIPS ciphers on the shared Gateway, catalog items tiered by FIPS compatibility, and
cs scan fipsto verify it end to end. - Private access.
--var enable_vpn=trueadds an OpenVPN host with its own PKI (cs vpn add-user|users|revoke| connect|disconnect) or a Tailscale subnet router. - Day-2 operations.
cs update-ip,cs provision,cs status|output|inventory, a per-environment audit trail
and redacted logs,cs troubleshootwith known failure signatures and fixes, andcs undowith fifteen undo points
per environment and fifteen for global settings (at most five of one kind). - FinOps.
cs finops estimate(offline, from the environment's own inventory),cs finops cloud(AWS Cost
Explorer, Azure Cost Management),cs finops k8s(OpenCost allocation) andcs finops report. - Managed data services.
cs databricksandcs snowflakewith per-environment connection profiles. - Web console.
cs enable uistarts a local, token-protected console on127.0.0.1:7434with a setup wizard,
environment actions, the platform catalog, DR, chaos, scans, reports, agents, MCP, the credential vault and the
help pages. Every action runs the samecloudseedcommand and streams its output. - MCP server.
cs setup mcpexposes every feature as an MCP tool (30 tools, plus resources and prompts) over
Streamable HTTP on loopback with a bearer token, or stdio, and connects Claude Code, Claude Desktop, Codex, Cursor,
Windsurf, Gemini CLI and VS Code. Destructive tools requireconfirm=true. - Agentic mode.
cs agentic "<task>"with the built-in agent (Claude API) or Claude Code, Codex, Gemini or Grok;
a deterministic "headliner" research brief; bundled Agent Skills; credentials stripped from the agent and served to
child commands over a per-session socket; output redacted; human-only commands refused in agent sessions. - Explain everything.
cs explain <feature|target|command|item>shows how each capability is implemented, which
files and cloud resources it uses and where its state lives.cs helpcovers every command, topic, variable and
output. - Dependencies your way. Install tools locally (
cs deps install, SHA-256-verified official releases or
Homebrew), run everything in a Docker or Podman image (--runtime container), or build a single binary
(cs deps bundle). - Credential vault.
cs creds set|list|unset|clearkeeps cloud keys and API tokens in a local 0600 file used by
every command, the console and the MCP server. - Project. Apache-2.0 license, documentation site with a quick start and 15 step-by-step scenarios, contribution
guide, security policy and CI: unit tests on Python 3.9 and 3.12 on Linux and macOS,terraform fmt,validateand
the mockedterraform testsuites,go vet/test/buildfor the VMware provider, and the scenario scripts in
dry-run mode.
Docs: https://nimeshbuilds.github.io/cloudseed/ · Quick start: https://nimeshbuilds.github.io/cloudseed/getting-started/quickstart/ · 15 scenarios: https://nimeshbuilds.github.io/cloudseed/scenarios/