feat(cli): validate prerequisites and seal installation plans

This commit is contained in:
Codex
2026-09-28 17:03:25 +02:00
parent b9c3369e7b
commit 55f3569e55
30 changed files with 2271 additions and 4 deletions
+122
View File
@@ -0,0 +1,122 @@
// Package preflight checks a prepared installation without creating application state.
package preflight
import (
"context"
"encoding/json"
"io"
"os"
"runtime"
"strconv"
"strings"
"time"
"github.com/aritmolab/thothii/tools/tht/internal/compose"
)
const Protocol = 1
type Check struct {
ID string `json:"id"`
Outcome string `json:"outcome"`
Field string `json:"field"`
Action string `json:"action"`
}
type Report struct {
SchemaVersion int `json:"schema_version"`
OK bool `json:"ok"`
Checks []Check `json:"checks"`
}
func NewReport() Report { return Report{SchemaVersion: 1, OK: true, Checks: []Check{}} }
func (r *Report) Add(id, outcome, field, action string) {
r.Checks = append(r.Checks, Check{id, outcome, field, action})
if outcome == "error" {
r.OK = false
}
}
func (r *Report) Merge(other Report) {
r.Checks = append(r.Checks, other.Checks...)
r.OK = r.OK && other.OK
}
func (r Report) JSON() string { data, _ := json.Marshal(r); return string(data) }
type Runner interface {
Run(context.Context, []string, io.Reader) (compose.Result, error)
}
func docker(ctx context.Context, runner Runner, args ...string) (compose.Result, error) {
bound, cancel := context.WithTimeout(ctx, 15*time.Second)
defer cancel()
return runner.Run(bound, args, nil)
}
type Requirements struct {
CPUs int `json:"cpus"`
MemoryBytes uint64 `json:"memory_bytes"`
DiskBytes uint64 `json:"disk_bytes"`
}
func DefaultRequirements() Requirements { return Requirements{2, 4 << 30, 10 << 30} }
type Host struct {
OS string
Arch string
Kernel string
Distribution string
FreeBytes uint64
}
func InspectHost(directory string) (Host, error) {
h := Host{OS: runtime.GOOS, Arch: runtime.GOARCH}
kernel, _ := os.ReadFile("/proc/sys/kernel/osrelease")
h.Kernel = strings.ToLower(string(kernel))
distribution, _ := os.ReadFile("/etc/os-release")
h.Distribution = strings.ToLower(string(distribution))
var err error
h.FreeBytes, err = freeBytes(directory)
return h, err
}
func CheckHost(ctx context.Context, runner Runner, host Host, minimum Requirements) Report {
r := NewReport()
check := func(id string, ok bool, action string) {
outcome := "passed"
if !ok {
outcome = "error"
}
r.Add(id, outcome, "host", action)
}
check("host-platform", (host.OS == "linux" || host.OS == "darwin") && (host.Arch == "amd64" || host.Arch == "arm64"), "Use the Linux executable in Ubuntu WSL2/Omarchy, or the matching macOS executable; native Windows is not the installation path.")
if strings.Contains(strings.ToLower(host.Kernel), "microsoft") {
check("wsl2", strings.Contains(strings.ToLower(host.Kernel), "wsl2") && strings.Contains(host.Distribution, "ubuntu"), "Use Ubuntu WSL2 and enable Docker Desktop integration for that distribution.")
}
info, err := docker(ctx, runner, "info", "--format", "{{json .}}")
var parsed struct {
OSType string
Architecture string
NCPU int
MemTotal uint64
}
good := err == nil && json.Unmarshal([]byte(info.Stdout), &parsed) == nil
check("docker-daemon", good && parsed.OSType == "linux", "Start a reachable Linux Docker daemon for this user.")
arch := parsed.Architecture
if arch == "x86_64" {
arch = "amd64"
}
if arch == "aarch64" {
arch = "arm64"
}
check("docker-architecture", good && arch == host.Arch, "Use a Linux Docker daemon matching the host architecture; emulation is not certified.")
check("docker-resources", good && parsed.NCPU >= minimum.CPUs && parsed.MemTotal >= minimum.MemoryBytes, "Allocate at least the release CPU and memory minimum to Docker.")
check("installation-disk", host.FreeBytes >= minimum.DiskBytes && host.FreeBytes > 0, "Provide the release minimum free space on the installation filesystem.")
result, err := docker(ctx, runner, "compose", "version", "--short")
parts := strings.Split(strings.TrimPrefix(strings.TrimSpace(result.Stdout), "v"), ".")
major, _ := strconv.Atoi(parts[0])
minor := 0
if len(parts) > 1 {
minor, _ = strconv.Atoi(parts[1])
}
check("docker-compose", err == nil && (major > 2 || (major == 2 && minor >= 24)), "Install Docker Compose v2.24 or newer.")
r.Add("daemon-storage", "warning", "host", "Host disk capacity does not measure a Docker Desktop VM disk; reserve equivalent Docker storage and verify volume allocation during setup.")
return r
}
+60
View File
@@ -0,0 +1,60 @@
package preflight
import (
"context"
"encoding/json"
"os"
"path/filepath"
"github.com/aritmolab/thothii/tools/tht/internal/config"
)
// CheckCompose resolves the distributed release plus authored overrides without starting services.
func CheckCompose(ctx context.Context, runner Runner, installation config.Installation, m Manifest, manifestPath, platform string) Report {
r := NewReport()
args := []string{"compose", "--project-directory", installation.ProjectDirectory, "--env-file", installation.EnvFile}
for _, name := range m.Compose {
args = append(args, "-f", filepath.Join(filepath.Dir(manifestPath), filepath.FromSlash(name)))
}
for _, path := range installation.Overrides {
if _, err := os.Stat(path); os.IsNotExist(err) {
// Only the two well-known distribution-owned transport overlays can be relocated.
name := "deploy/" + filepath.Base(path)
if path != filepath.Join(installation.ProjectDirectory, filepath.FromSlash(name)) || (name != "deploy/compose.git-https.yaml" && name != "deploy/compose.git-ssh.yaml") || m.Files[name] == "" {
r.Add("compose-assets", "error", "overrides", "Include the selected Git transport overlay in the verified release.")
return r
}
path = filepath.Join(filepath.Dir(manifestPath), filepath.FromSlash(name))
}
args = append(args, "-f", path)
}
args = append(args, "config", "--format", "json")
result, err := docker(ctx, runner, args...)
var effective struct {
Services map[string]struct {
Image string `json:"image"`
Build any `json:"build"`
Platform string `json:"platform"`
} `json:"services"`
}
if err != nil || json.Unmarshal([]byte(result.Stdout), &effective) != nil {
r.Add("compose-configuration", "error", "operator.env", "Correct the effective Compose configuration using the release assets and prepared environment; no raw output is logged.")
return r
}
roles := map[string]string{"core": "core", "catalog-migrate": "core", "workspace-maintenance": "core", "frontend": "frontend", "catalog-db": "catalog", "qdrant": "qdrant", "embedding": "embedding", "embedding-model-init": "embedding"}
for service, role := range roles {
entry, exists := effective.Services[service]
if !exists || entry.Build != nil || entry.Image != m.Images[role][platform] || (entry.Platform != "" && entry.Platform != platform) {
r.Add("compose-image-"+service, "error", "release.compose", "Each runtime and maintenance service must use its released immutable image without a source build.")
}
}
for service := range effective.Services {
if _, ok := roles[service]; !ok {
r.Add("compose-service", "error", "release.compose", "Additional services need an explicit release contract before execution.")
}
}
if r.OK {
r.Add("compose-configuration", "passed", "release.compose", "Effective Compose configuration uses the complete released image set.")
}
return r
}
+13
View File
@@ -0,0 +1,13 @@
//go:build !windows
package preflight
import "golang.org/x/sys/unix"
func freeBytes(path string) (uint64, error) {
var stat unix.Statfs_t
if err := unix.Statfs(path, &stat); err != nil {
return 0, err
}
return uint64(stat.Bavail) * uint64(stat.Bsize), nil
}
@@ -0,0 +1,15 @@
//go:build windows
package preflight
import "golang.org/x/sys/windows"
func freeBytes(path string) (uint64, error) {
p, err := windows.UTF16PtrFromString(path)
if err != nil {
return 0, err
}
var available uint64
err = windows.GetDiskFreeSpaceEx(p, &available, nil, nil)
return available, err
}
+216
View File
@@ -0,0 +1,216 @@
package preflight
import (
"context"
"crypto/tls"
"crypto/x509"
"errors"
"io"
"net"
"net/http"
"net/url"
"strconv"
"strings"
"time"
"github.com/aritmolab/thothii/tools/tht/internal/config"
"github.com/aritmolab/thothii/tools/tht/internal/safeio"
"golang.org/x/crypto/ssh"
"golang.org/x/crypto/ssh/knownhosts"
)
func CheckExternal(ctx context.Context, installation config.Installation) Report {
r := NewReport()
value := func(name string) string { result, _ := installation.EnvironmentValue(name); return result }
err := checkGit(ctx, installation, value)
outcome := "passed"
if err != nil {
outcome = "error"
}
r.Add("workspace-remote", outcome, "workspaceRepository", "Require authenticated read access to the configured Git remote and branch using the prepared trust/credential files.")
for name, provider := range installation.ModelCatalog.Providers {
if provider.Endpoint == nil {
r.Add("provider-"+name, "warning", "modelCatalog.providers", "Built-in provider endpoint resolution belongs to the bundled Pi SDK. The required Pi runtime smoke check verifies model availability and credentials; preflight makes no billable generation requests.")
continue
}
parsed, err := url.Parse(provider.Endpoint.BaseURL)
if err == nil {
err = probeOrigin(ctx, parsed)
}
outcome := "passed"
if err != nil {
outcome = "error"
}
r.Add("provider-"+name, outcome, "modelCatalog.providers."+name+".endpoint", "Require DNS/TCP/TLS reachability of the configured provider origin. Credential/model eligibility still requires the bundled Pi runtime smoke check; no generation request is sent here.")
}
return r
}
func probeOrigin(ctx context.Context, target *url.URL) error {
bound, cancel := context.WithTimeout(ctx, 5*time.Second)
defer cancel()
port := target.Port()
if port == "" {
port = "443"
if target.Scheme == "http" {
port = "80"
}
}
address := net.JoinHostPort(target.Hostname(), port)
var connection net.Conn
var err error
if target.Scheme == "https" {
dialer := tls.Dialer{NetDialer: &net.Dialer{Timeout: 5 * time.Second}, Config: &tls.Config{MinVersion: tls.VersionTLS12, ServerName: target.Hostname()}}
connection, err = dialer.DialContext(bound, "tcp", address)
} else {
connection, err = (&net.Dialer{Timeout: 5 * time.Second}).DialContext(bound, "tcp", address)
}
if err == nil {
connection.Close()
}
return err
}
func checkGit(ctx context.Context, installation config.Installation, value func(string) string) error {
remote := installation.WorkspaceRepository.Remote
if installation.WorkspaceRepository.Access == "ssh" && strings.HasPrefix(remote, "git@") {
host, path, found := strings.Cut(strings.TrimPrefix(remote, "git@"), ":")
if !found {
return errors.New("invalid Git remote")
}
return checkSSHGit(ctx, &url.URL{Scheme: "ssh", User: url.User("git"), Host: host, Path: path}, installation.WorkspaceRepository.Branch, value)
}
u, err := url.Parse(remote)
if err != nil {
return errors.New("remote unavailable")
}
if installation.WorkspaceRepository.Access == "ssh" {
return checkSSHGit(ctx, u, installation.WorkspaceRepository.Branch, value)
}
ca, err := safeio.ReadCanonicalPrivateRegular(value("THT_WORKSPACE_GIT_CA_FILE"), 64<<10)
if err != nil {
return err
}
pool, err := x509.SystemCertPool()
if err != nil {
pool = x509.NewCertPool()
}
if !pool.AppendCertsFromPEM(ca) {
return errors.New("invalid Git trust")
}
credentials, err := safeio.ReadCanonicalPrivateRegular(value("THT_WORKSPACE_GIT_CREDENTIALS_FILE"), 64<<10)
if err != nil {
return err
}
var user, password string
for _, line := range strings.Split(string(credentials), "\n") {
if strings.TrimSpace(line) == "" {
continue
}
credential, err := url.Parse(strings.TrimSpace(line))
if err != nil || credential.User == nil {
return errors.New("invalid Git credentials")
}
if credential.Scheme == u.Scheme && credential.Host == u.Host && (credential.Path == "" || credential.Path == u.Path) {
user = credential.User.Username()
password, _ = credential.User.Password()
}
}
u.Path = strings.TrimSuffix(u.Path, "/") + "/info/refs"
u.RawQuery = "service=git-upload-pack"
bound, cancel := context.WithTimeout(ctx, 5*time.Second)
defer cancel()
request, err := http.NewRequestWithContext(bound, http.MethodGet, u.String(), nil)
if err != nil {
return err
}
if user != "" {
request.SetBasicAuth(user, password)
}
transport := &http.Transport{TLSClientConfig: &tls.Config{RootCAs: pool, MinVersion: tls.VersionTLS12}, Proxy: http.ProxyFromEnvironment}
defer transport.CloseIdleConnections()
client := http.Client{Transport: transport, Timeout: 5 * time.Second, CheckRedirect: func(*http.Request, []*http.Request) error { return http.ErrUseLastResponse }}
response, err := client.Do(request)
if err != nil {
return err
}
defer response.Body.Close()
if response.StatusCode != http.StatusOK {
return errors.New("Git remote refused")
}
data, err := io.ReadAll(io.LimitReader(response.Body, (1<<20)+1))
if err != nil || len(data) > 1<<20 || !advertisesBranch(data, installation.WorkspaceRepository.Branch) {
return errors.New("Git branch unavailable")
}
return nil
}
func advertisesBranch(data []byte, branch string) bool {
for _, line := range strings.Split(string(data), "\n") {
line = strings.SplitN(line, "\x00", 2)[0]
if strings.HasSuffix(line, " refs/heads/"+branch) {
return true
}
}
return false
}
func checkSSHGit(ctx context.Context, target *url.URL, branch string, value func(string) string) error {
// Git paths and branch names are already validated by the canonical installation loader.
if target.Scheme != "ssh" || target.User == nil || strings.ContainsAny(target.Path, "'\r\n\x00") {
return errors.New("use canonical ssh:// remote")
}
key, err := safeio.ReadCanonicalPrivateRegular(value("THT_WORKSPACE_GIT_SSH_KEY_FILE"), 64<<10)
if err != nil {
return err
}
signer, err := ssh.ParsePrivateKey(key)
if err != nil {
return err
}
hostKey, err := knownhosts.New(value("THT_WORKSPACE_GIT_KNOWN_HOSTS_FILE"))
if err != nil {
return err
}
port := target.Port()
if port == "" {
port = "22"
}
if _, err := strconv.Atoi(port); err != nil {
return err
}
address := net.JoinHostPort(target.Hostname(), port)
bound, cancel := context.WithTimeout(ctx, 5*time.Second)
defer cancel()
connection, err := (&net.Dialer{Timeout: 5 * time.Second}).DialContext(bound, "tcp", address)
if err != nil {
return err
}
defer connection.Close()
deadline, _ := bound.Deadline()
_ = connection.SetDeadline(deadline)
clientConnection, channels, requests, err := ssh.NewClientConn(connection, address, &ssh.ClientConfig{User: target.User.Username(), Auth: []ssh.AuthMethod{ssh.PublicKeys(signer)}, HostKeyCallback: hostKey, Timeout: 5 * time.Second})
if err != nil {
return err
}
client := ssh.NewClient(clientConnection, channels, requests)
defer client.Close()
session, err := client.NewSession()
if err != nil {
return err
}
defer session.Close()
stdout, err := session.StdoutPipe()
if err != nil {
return err
}
stdin, err := session.StdinPipe()
if err != nil {
return err
}
if err = session.Start("git-upload-pack --advertise-refs '" + target.Path + "'"); err != nil {
return err
}
_ = stdin.Close()
data, err := io.ReadAll(io.LimitReader(stdout, (1<<20)+1))
if err != nil || len(data) > 1<<20 || !advertisesBranch(data, branch) {
return errors.New("Git branch unavailable")
}
return nil
}
@@ -0,0 +1,62 @@
package preflight
import (
"context"
"encoding/pem"
"net/http"
"net/http/httptest"
"os"
"path/filepath"
"strings"
"testing"
"github.com/aritmolab/thothii/tools/tht/internal/config"
)
func TestGitProbeAuthenticatesAndRejectsWrongBranchAndCredentials(t *testing.T) {
status := 200
server := httptest.NewTLSServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
user, password, ok := r.BasicAuth()
if !ok || user != "reader" || password != "PRIVATE_SENTINEL" {
w.WriteHeader(401)
return
}
if r.Method != "GET" || r.URL.Path != "/workspaces.git/info/refs" || r.URL.RawQuery != "service=git-upload-pack" {
t.Error("unexpected Git mutation/request")
w.WriteHeader(400)
return
}
w.WriteHeader(status)
_, _ = w.Write([]byte("0044" + strings.Repeat("a", 40) + " refs/heads/main\n"))
}))
defer server.Close()
root, _ := filepath.EvalSymlinks(t.TempDir())
ca := filepath.Join(root, "ca.pem")
credentials := filepath.Join(root, "credentials")
_ = os.WriteFile(ca, pem.EncodeToMemory(&pem.Block{Type: "CERTIFICATE", Bytes: server.Certificate().Raw}), 0o600)
_ = os.WriteFile(credentials, []byte(strings.Replace(server.URL, "https://", "https://reader:PRIVATE_SENTINEL@", 1)), 0o600)
installation := config.Installation{WorkspaceRepository: config.WorkspaceRepository{Remote: server.URL + "/workspaces.git", Branch: "main", Access: "https"}}
value := func(name string) string {
if name == "THT_WORKSPACE_GIT_CA_FILE" {
return ca
}
return credentials
}
if err := checkGit(context.Background(), installation, value); err != nil {
t.Fatal(err)
}
installation.WorkspaceRepository.Branch = "missing"
if checkGit(context.Background(), installation, value) == nil {
t.Fatal("missing branch accepted")
}
installation.WorkspaceRepository.Branch = "main"
status = 503
if checkGit(context.Background(), installation, value) == nil {
t.Fatal("unavailable existing Git accepted")
}
status = 200
_ = os.WriteFile(credentials, []byte(strings.Replace(server.URL, "https://", "https://reader:rotated@", 1)), 0o600)
if checkGit(context.Background(), installation, value) == nil {
t.Fatal("bad credential accepted")
}
}
+287
View File
@@ -0,0 +1,287 @@
package preflight
import (
"crypto/hmac"
"crypto/rand"
"crypto/sha256"
"encoding/hex"
"encoding/json"
"errors"
"io/fs"
"os"
"path/filepath"
"slices"
"strings"
"github.com/aritmolab/thothii/tools/tht/internal/config"
"github.com/aritmolab/thothii/tools/tht/internal/safeio"
"github.com/aritmolab/thothii/tools/tht/internal/version"
"gopkg.in/yaml.v3"
)
type Plan struct {
SchemaVersion int `json:"schema_version"`
ValidatorProtocol int `json:"validator_protocol"`
Validator version.Info `json:"validator"`
Installation config.Installation `json:"installation"`
Release Manifest `json:"release"`
Platform string `json:"platform"`
WorkspaceDirectory string `json:"workspace_directory"`
WorkspaceRevision string `json:"workspace_revision"`
Inputs []string `json:"inputs"`
AbsentInputs []string `json:"absent_inputs"`
InputSeal string `json:"input_seal"`
Report Report `json:"report"`
}
func invalidPlan() error {
return errors.New("plan inputs changed or protected plan files are unavailable; repeat validation and produce a new plan")
}
// CaptureInputs gives a value-free freshness guard around live probes and document validation.
func CaptureInputs(paths []string, workspace string, absent ...string) (func() bool, error) {
key := make([]byte, 32)
if _, err := rand.Read(key); err != nil {
return nil, invalidPlan()
}
probe := Plan{Inputs: paths, AbsentInputs: absent}
before, err := seal(probe, key)
if err != nil {
return nil, err
}
files, err := treeFiles(workspace, true)
if err != nil {
return nil, err
}
return func() bool {
after, err := seal(probe, key)
current, treeErr := treeFiles(workspace, true)
return err == nil && treeErr == nil && hmac.Equal([]byte(before), []byte(after)) && slices.Equal(files, current)
}, nil
}
// A separate owner-only random key prevents public/offline guessing of low-entropy secrets.
// Live prerequisites must still be rechecked immediately before any execution or resumption.
func seal(plan Plan, key []byte) (string, error) {
for _, path := range plan.AbsentInputs {
if _, err := os.Lstat(path); !errors.Is(err, os.ErrNotExist) {
return "", invalidPlan()
}
}
mac := hmac.New(sha256.New, key)
plan.InputSeal = ""
data, err := json.Marshal(plan)
if err != nil {
return "", invalidPlan()
}
mac.Write(data)
var total int
for _, path := range plan.Inputs {
contents, err := safeio.ReadCanonicalRegular(path, 32<<20)
if err != nil {
return "", invalidPlan()
}
total += len(contents)
if total > 256<<20 {
return "", invalidPlan()
}
length, _ := json.Marshal([]any{path, len(contents)})
mac.Write(length)
mac.Write(contents)
}
return hex.EncodeToString(mac.Sum(nil)), nil
}
func AbsentOverrides(installation config.Installation) []string {
paths := []string{}
for _, path := range installation.Overrides {
if _, err := os.Lstat(path); errors.Is(err, os.ErrNotExist) {
paths = append(paths, path)
}
}
return paths
}
func WritePlan(path string, plan *Plan, guards ...func() bool) error {
if !plan.Report.OK || plan.ValidatorProtocol != Protocol || !filepath.IsAbs(path) {
return invalidPlan()
}
if exists, err := safeio.PreflightPrivateDirectory(filepath.Dir(path)); err != nil || !exists {
return invalidPlan()
}
for _, target := range []string{path, path + ".key"} {
if _, err := os.Lstat(target); !errors.Is(err, os.ErrNotExist) {
return invalidPlan()
}
}
key := make([]byte, 32)
if _, err := rand.Read(key); err != nil {
return invalidPlan()
}
var err error
plan.InputSeal, err = seal(*plan, key)
if err != nil {
return err
}
for _, guard := range guards {
if guard == nil || !guard() {
return invalidPlan()
}
}
data, _ := json.MarshalIndent(plan, "", " ")
if err := safeio.WriteCanonicalNewPrivateFile(path+".key", key, 0o600); err != nil {
return invalidPlan()
}
if err := safeio.WriteCanonicalNewPrivateFile(path, append(data, '\n'), 0o600); err != nil {
_ = safeio.RemoveCanonicalPrivateRegular(path + ".key")
return invalidPlan()
}
return nil
}
func VerifyPlanInputs(path string) error {
data, err := safeio.ReadCanonicalPrivateRegular(path, 4<<20)
if err != nil {
return invalidPlan()
}
key, err := safeio.ReadCanonicalPrivateRegular(path+".key", 32)
if err != nil || len(key) != 32 {
return invalidPlan()
}
var plan Plan
if json.Unmarshal(data, &plan) != nil || plan.SchemaVersion != 1 || plan.ValidatorProtocol != Protocol || !plan.Report.OK || len(plan.Inputs) == 0 || len(plan.Inputs) > 10000 {
return invalidPlan()
}
actual, err := seal(plan, key)
if err != nil || !hmac.Equal([]byte(actual), []byte(plan.InputSeal)) {
return invalidPlan()
}
// Detect added or removed workspace files as well as changes to known file bytes.
if plan.WorkspaceDirectory != "" {
files, err := treeFiles(plan.WorkspaceDirectory, true)
if err != nil {
return invalidPlan()
}
for _, file := range files {
if !slices.Contains(plan.Inputs, file) {
return invalidPlan()
}
}
}
return nil
}
func treeFiles(root string, skipGit bool) ([]string, error) {
paths := []string{}
count := 0
err := filepath.WalkDir(root, func(path string, entry fs.DirEntry, err error) error {
if err != nil {
return invalidPlan()
}
count++
if count > 10000 {
return invalidPlan()
}
if skipGit && entry.Name() == ".git" {
if entry.IsDir() {
return filepath.SkipDir
}
return nil
}
if entry.IsDir() {
return nil
}
if !entry.Type().IsRegular() {
return invalidPlan()
}
paths = append(paths, path)
return nil
})
return paths, err
}
// CollectInputs fingerprints exact prepared contents, with normalized configuration in Plan.
// Referenced credentials are sealed, never copied. Git object stores are excluded.
func CollectInputs(installation config.Installation, workspace, bootstrap, manifest string) ([]string, string, error) {
paths := []string{installation.Path, installation.EnvFile, bootstrap, manifest}
for _, root := range []string{workspace, installation.AuthenticationDirectory()} {
files, err := treeFiles(root, root == workspace)
if err != nil {
return nil, "", err
}
paths = append(paths, files...)
}
secrets, err := installation.SecretFiles()
if err != nil {
return nil, "", invalidPlan()
}
for _, path := range secrets {
paths = append(paths, path)
}
data, err := safeio.ReadCanonicalPrivateRegular(bootstrap, 1<<20)
if err != nil {
return nil, "", invalidPlan()
}
var bindings struct {
Databases []struct {
SecretFiles map[string]string `yaml:"secretFiles"`
EvidenceSecretFiles map[string]string `yaml:"evidenceSecretFiles"`
} `yaml:"databases"`
}
if yaml.Unmarshal(data, &bindings) != nil {
return nil, "", invalidPlan()
}
for _, entry := range bindings.Databases {
for _, values := range []map[string]string{entry.SecretFiles, entry.EvidenceSecretFiles} {
for _, path := range values {
paths = append(paths, path)
}
}
}
m, err := LoadManifest(manifest)
if err != nil {
return nil, "", err
}
for name := range m.Files {
paths = append(paths, filepath.Join(filepath.Dir(manifest), filepath.FromSlash(name)))
}
for _, path := range installation.Overrides {
if _, err := os.Lstat(path); err == nil {
paths = append(paths, path)
}
}
revision := "content-snapshot"
head := filepath.Join(workspace, ".git", "HEAD")
if data, err := safeio.ReadCanonicalRegular(head, 1024); err == nil {
paths = append(paths, head)
value := strings.TrimSpace(string(data))
if strings.HasPrefix(value, "ref: refs/") {
ref := strings.TrimPrefix(value, "ref: ")
if safeRelative(ref) {
path := filepath.Join(workspace, ".git", filepath.FromSlash(ref))
if data, err := safeio.ReadCanonicalRegular(path, 1024); err == nil {
paths = append(paths, path)
value = strings.TrimSpace(string(data))
}
}
}
if len(value) == 40 {
if _, err := hex.DecodeString(value); err == nil {
revision = value
}
}
}
slices.Sort(paths)
paths = slices.Compact(paths)
return paths, revision, nil
}
func AddRuntimeObligations(report *Report) {
for _, check := range []Check{
{"container-network", "deferred-to-runtime", "bindings", "Repeat authenticated DWH and external connectivity checks from the core network."},
{"catalog-initialization", "deferred-to-runtime", "catalog", "Apply migrations and verify Catalog health plus prepared binding import."},
{"pi-operation", "deferred-to-runtime", "release.components.pi", "Verify the bundled Pi version and authenticated provider/model smoke operation inside core; do not install Pi on the host."},
{"local-embedding", "deferred-to-runtime", "modelCatalog.embedding", "Initialize the local embedding model and verify returned vector dimensions."},
{"workspace-preprocessing", "deferred-to-runtime", "workspaces", "Sync the exact verified workspace contents, materialize Evidence, preprocess and verify collections."},
{"workspace-readiness", "deferred-to-runtime", "workspaces", "Complete required administrative and human review gates before claiming final readiness."},
} {
report.Checks = append(report.Checks, check)
}
}
+68
View File
@@ -0,0 +1,68 @@
package preflight
import (
"github.com/aritmolab/thothii/tools/tht/internal/safeio"
"os"
"path/filepath"
"strings"
"testing"
)
func TestPlanBindsInputsAndDetectsCredentialRotationWithoutPublicSecretHashes(t *testing.T) {
root, _ := filepath.EvalSymlinks(t.TempDir())
if err := safeio.ProtectPrivateDirectory(root); err != nil {
t.Fatal(err)
}
input := filepath.Join(root, "descriptor.yaml")
secret := filepath.Join(root, "credential")
for path, data := range map[string]string{input: "schemaVersion: 2\n", secret: "PRIVATE_SENTINEL"} {
if err := safeio.WriteCanonicalNewPrivateFile(path, []byte(data), 0o600); err != nil {
t.Fatal(err)
}
}
plan := Plan{SchemaVersion: 1, ValidatorProtocol: Protocol, Inputs: []string{input, secret}, WorkspaceRevision: "content-snapshot", Report: NewReport()}
output := filepath.Join(root, "plan.json")
if err := WritePlan(output, &plan); err != nil {
t.Fatal(err)
}
data, _ := os.ReadFile(output)
if strings.Contains(string(data), "PRIVATE_SENTINEL") {
t.Fatal("secret in plan")
}
if err := VerifyPlanInputs(output); err != nil {
t.Fatal(err)
}
if err := WritePlan(output, &plan); err == nil {
t.Fatal("existing plan replaced")
}
if err := os.WriteFile(secret, []byte("rotated"), 0o600); err != nil {
t.Fatal(err)
}
if err := VerifyPlanInputs(output); err == nil {
t.Fatal("credential rotation did not invalidate plan")
}
if err := os.WriteFile(secret, []byte("PRIVATE_SENTINEL"), 0o600); err != nil {
t.Fatal(err)
}
if err := os.WriteFile(input, []byte("schemaVersion: 3\n"), 0o600); err != nil {
t.Fatal(err)
}
if err := VerifyPlanInputs(output); err == nil {
t.Fatal("document change did not invalidate plan")
}
if err := os.WriteFile(input, []byte("schemaVersion: 2\n"), 0o600); err != nil {
t.Fatal(err)
}
absent := filepath.Join(root, "transport-override.yaml")
plan.AbsentInputs = []string{absent}
second := filepath.Join(root, "second-plan.json")
if err := WritePlan(second, &plan); err != nil {
t.Fatal(err)
}
if err := safeio.WriteCanonicalNewPrivateFile(absent, []byte("services: {}\n"), 0o600); err != nil {
t.Fatal(err)
}
if VerifyPlanInputs(second) == nil {
t.Fatal("newly appearing override did not invalidate plan")
}
}
@@ -0,0 +1,118 @@
package preflight
import (
"context"
"encoding/json"
"errors"
"io"
"strings"
"testing"
"github.com/aritmolab/thothii/tools/tht/internal/compose"
"github.com/aritmolab/thothii/tools/tht/internal/config"
)
type fakeDocker struct {
calls [][]string
fail string
effective string
}
func (f *fakeDocker) Run(_ context.Context, args []string, _ io.Reader) (compose.Result, error) {
f.calls = append(f.calls, args)
if strings.Contains(strings.Join(args, " "), f.fail) && f.fail != "" {
return compose.Result{Stderr: "PRIVATE_SENTINEL"}, errors.New("PRIVATE_SENTINEL")
}
switch args[0] {
case "info":
return compose.Result{Stdout: `{"OSType":"linux","Architecture":"x86_64","NCPU":4,"MemTotal":17179869184}`}, nil
case "compose":
if args[len(args)-1] == "json" {
if f.effective != "" {
return compose.Result{Stdout: f.effective}, nil
}
return compose.Result{Stdout: `{"services":{"core":{"image":"example/core:latest"}}}`}, nil
}
return compose.Result{Stdout: "2.39.0"}, nil
case "manifest":
return compose.Result{Stdout: `{"Descriptor":{"digest":"sha256:` + strings.Repeat("a", 64) + `","platform":{"os":"linux","architecture":"amd64"}}}`}, nil
}
return compose.Result{}, errors.New("unexpected command")
}
func TestComposeRejectsIncompleteMutableServiceSet(t *testing.T) {
r := CheckCompose(context.Background(), &fakeDocker{}, config.Installation{ProjectDirectory: "/private", EnvFile: "/private/operator.env"}, Manifest{Compose: []string{"compose.yaml"}}, "/release/manifest.json", "linux/amd64")
if r.OK {
t.Fatal("unreleased service set accepted")
}
}
func TestComposeRejectsPlatformOverrideAgainstSelectedImage(t *testing.T) {
m := Manifest{Images: map[string]map[string]string{}, Compose: []string{"compose.yaml"}}
services := map[string]map[string]string{}
for service, role := range map[string]string{"core": "core", "frontend": "frontend", "catalog-db": "catalog", "catalog-migrate": "core", "workspace-maintenance": "core", "qdrant": "qdrant", "embedding": "embedding", "embedding-model-init": "embedding"} {
m.Images[role] = map[string]string{"linux/amd64": "docker.io/example/" + role + "@sha256:" + strings.Repeat("a", 64)}
services[service] = map[string]string{"image": m.Images[role]["linux/amd64"]}
}
services["core"]["platform"] = "linux/arm64"
data, _ := json.Marshal(map[string]any{"services": services})
r := CheckCompose(context.Background(), &fakeDocker{effective: string(data)}, config.Installation{}, m, "/release/manifest.json", "linux/amd64")
if r.OK {
t.Fatal("incompatible Compose platform accepted")
}
}
func TestHostChecksAreReadOnlyAndRejectUnavailableDocker(t *testing.T) {
f := &fakeDocker{}
host := Host{OS: "linux", Arch: "amd64", Kernel: "6.6-microsoft-standard-WSL2", Distribution: "ubuntu", FreeBytes: 30 << 30}
r := CheckHost(context.Background(), f, host, Requirements{CPUs: 2, MemoryBytes: 4 << 30, DiskBytes: 10 << 30})
if !r.OK {
t.Fatalf("host rejected: %+v", r)
}
for _, args := range f.calls {
if args[0] != "info" && !(args[0] == "compose" && args[1] == "version") {
t.Fatalf("mutating call: %v", args)
}
}
f.fail = "info"
r = CheckHost(context.Background(), f, host, Requirements{CPUs: 2, MemoryBytes: 4 << 30, DiskBytes: 10 << 30})
if r.OK || strings.Contains(r.JSON(), "PRIVATE_SENTINEL") {
t.Fatalf("unsafe success/report: %s", r.JSON())
}
host.Kernel = "4.4-microsoft"
if CheckHost(context.Background(), &fakeDocker{}, host, Requirements{CPUs: 2, MemoryBytes: 4 << 30, DiskBytes: 10 << 30}).OK {
t.Fatal("WSL1 accepted")
}
host.OS = "windows"
if CheckHost(context.Background(), &fakeDocker{}, host, Requirements{}).OK {
t.Fatal("native Windows accepted instead of WSL2")
}
}
func TestReleaseChecksEveryImmutableImageAndPlatform(t *testing.T) {
m := Manifest{SchemaVersion: 1, Version: "1.0.0", Revision: strings.Repeat("b", 40), ValidatorProtocol: 1, Requirements: Requirements{CPUs: 2, MemoryBytes: 4 << 30, DiskBytes: 10 << 30}, Components: []string{"pi", "catalog-migrations", "workspace-maintenance"}, Images: map[string]map[string]string{}}
for _, service := range []string{"core", "frontend", "catalog", "qdrant", "embedding"} {
m.Images[service] = map[string]string{"linux/amd64": "docker.io/example/" + service + "@sha256:" + strings.Repeat("a", 64)}
}
m.Files = map[string]string{"deploy/compose.yaml": strings.Repeat("c", 64)}
m.Compose = []string{"deploy/compose.yaml"}
if err := m.Validate(); err != nil {
t.Fatal(err)
}
f := &fakeDocker{}
r := CheckImages(context.Background(), f, m, "linux/amd64")
if !r.OK || len(f.calls) != 5 {
t.Fatalf("images not checked: %s calls=%d", r.JSON(), len(f.calls))
}
if CheckImages(context.Background(), f, m, "linux/arm64").OK {
t.Fatal("unsupported release architecture accepted")
}
f.fail = "frontend"
if CheckImages(context.Background(), f, m, "linux/amd64").OK {
t.Fatal("missing image accepted")
}
m.Images["core"]["linux/amd64"] = "example/core:latest"
if m.Validate() == nil {
t.Fatal("mutable tag accepted")
}
}
+131
View File
@@ -0,0 +1,131 @@
package preflight
import (
"bytes"
"context"
"crypto/sha256"
"encoding/hex"
"encoding/json"
"errors"
"io"
"path/filepath"
"regexp"
"slices"
"strings"
"github.com/aritmolab/thothii/tools/tht/internal/safeio"
)
// Manifest is the publication/consumer contract. Each platform maps to a single-image digest,
// not a mutable tag or multi-platform index. Maintenance roles use Images["core"].
type Manifest struct {
SchemaVersion int `json:"schema_version"`
Version string `json:"version"`
Revision string `json:"revision"`
ValidatorProtocol int `json:"validator_protocol"`
Requirements Requirements `json:"requirements"`
Components []string `json:"components"`
Images map[string]map[string]string `json:"images"`
Files map[string]string `json:"files"`
Compose []string `json:"compose"`
}
var hex256 = regexp.MustCompile(`^[a-f0-9]{64}$`)
var imageReference = regexp.MustCompile(`^docker\.io/[a-z0-9][a-z0-9._/-]*@sha256:[a-f0-9]{64}$`)
func invalidRelease() error {
return errors.New("release manifest or packaged files are incomplete, incompatible or invalid")
}
func (m Manifest) Validate() error {
if m.SchemaVersion != 1 || m.ValidatorProtocol != Protocol || !regexp.MustCompile(`^[0-9]+\.[0-9]+\.[0-9]+(?:-[A-Za-z0-9.-]+)?$`).MatchString(m.Version) || !regexp.MustCompile(`^[a-f0-9]{40}$`).MatchString(m.Revision) {
return invalidRelease()
}
if m.Requirements.CPUs < 2 || m.Requirements.MemoryBytes < 4<<30 || m.Requirements.DiskBytes < 10<<30 {
return invalidRelease()
}
for _, component := range []string{"pi", "catalog-migrations", "workspace-maintenance"} {
if !slices.Contains(m.Components, component) {
return invalidRelease()
}
}
if len(m.Images) != 5 || len(m.Files) == 0 || len(m.Files) > 256 || len(m.Compose) == 0 || len(m.Compose) > 8 {
return invalidRelease()
}
for _, service := range []string{"core", "frontend", "catalog", "qdrant", "embedding"} {
if len(m.Images[service]) == 0 {
return invalidRelease()
}
for platform, ref := range m.Images[service] {
if (platform != "linux/amd64" && platform != "linux/arm64") || !imageReference.MatchString(ref) {
return invalidRelease()
}
}
}
for name, digest := range m.Files {
if !safeRelative(name) || !hex256.MatchString(digest) {
return invalidRelease()
}
}
for _, name := range m.Compose {
if _, ok := m.Files[name]; !ok {
return invalidRelease()
}
}
return nil
}
func safeRelative(name string) bool {
return name != "" && !strings.Contains(name, "\\") && !strings.Contains(name, ":") && !strings.HasPrefix(name, "/") && filepath.ToSlash(filepath.Clean(name)) == name && name != ".." && !strings.HasPrefix(name, "../") && name != "."
}
func LoadManifest(path string) (Manifest, error) {
var m Manifest
contents, err := safeio.ReadCanonicalRegular(path, 1<<20)
if err != nil {
return m, invalidRelease()
}
decoder := json.NewDecoder(bytes.NewReader(contents))
decoder.DisallowUnknownFields()
if decoder.Decode(&m) != nil || decoder.Decode(new(any)) != io.EOF {
return Manifest{}, invalidRelease()
}
if err = m.Validate(); err != nil {
return Manifest{}, err
}
for name, digest := range m.Files {
contents, err := safeio.ReadCanonicalRegular(filepath.Join(filepath.Dir(path), filepath.FromSlash(name)), 32<<20)
if err != nil {
return Manifest{}, invalidRelease()
}
sum := sha256.Sum256(contents)
if hex.EncodeToString(sum[:]) != digest {
return Manifest{}, invalidRelease()
}
}
return m, nil
}
func CheckImages(ctx context.Context, runner Runner, m Manifest, platform string) Report {
r := NewReport()
for _, service := range []string{"core", "frontend", "catalog", "qdrant", "embedding"} {
ref := m.Images[service][platform]
if ref == "" {
r.Add("image-"+service, "error", "release.images."+service, "Publish the selected Linux architecture before producing an executable plan.")
continue
}
result, err := docker(ctx, runner, "manifest", "inspect", "--verbose", ref)
var manifest struct {
Descriptor struct {
Digest string `json:"digest"`
Platform struct {
OS string `json:"os"`
Architecture string `json:"architecture"`
} `json:"platform"`
} `json:"Descriptor"`
}
good := err == nil && json.Unmarshal([]byte(result.Stdout), &manifest) == nil && manifest.Descriptor.Platform.OS+"/"+manifest.Descriptor.Platform.Architecture == platform && strings.HasSuffix(ref, "@"+manifest.Descriptor.Digest) && manifest.Descriptor.Digest != ""
outcome := "passed"
if !good {
outcome = "error"
}
r.Add("image-"+service, outcome, "release.images."+service, "Require the pinned digest to be publicly readable in Docker Hub for the selected Linux architecture.")
}
return r
}