Files
ThothII/tools/tht/internal/preflight/plan.go

288 lines
9.2 KiB
Go

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)
}
}