e145d48f22
* feat: add plugin system with grpc runtime and native ui plugins - go plugin backends ship as release tarballs (per-platform binaries, checksums) that kandev installs from url/upload or filesystem sideload+sync, then spawns and supervises via hashicorp/go-plugin (unix socket / loopback tcp+automtls) - strict typed grpc contract (kandev.plugin.v1): DeliverEvent/InvokeTool/HandleWebhook plus a capability-gated host service (state, secrets, emit); author sdk in pkg/pluginsdk; packaging cli in cmd/plugin-pack - native frontend bundles: window.registerKandevPlugin + registry/host loader with dynamic routes, nav items, slots and ws handlers; plugins settings page with install dialog, sync button and enable/disable/uninstall - gated behind the plugins feature flag (prod off, dev/e2e on); covered by unit and integration tests plus playwright e2e against a real packaged fixture plugin Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> * fix(plugins): hydrate base app context on plugin routes Plugin-owned SPA routes classify as RouteUnknown, so the boot payload shipped only feature flags — no workspaces, workflows, or kanban context. Native plugin UI (e.g. host.ui.TaskCreateDialog) then had an empty store. Hydrate the same base context as the Home route for unknown SPA routes. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> * feat(plugins): expose TaskCreateDialog on host.ui Lets a native plugin open kandev's real create-task modal (prefilled via initialValues) instead of POSTing directly. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> * fix(plugins): harden filesystem/URL sinks against CodeQL findings Add sink-level guards on the plugin trust boundary: securejoin containment on archive extraction (zip-slip), plugin-id validation before every fs_store path, http/https scheme validation on install-from-URL, and non-interpolated ws-handler error logging. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> * fix(plugins): address review findings (races, resource caps, lifecycle) - serialize per-plugin startup to close a TOCTOU double-spawn race, and let a plugin restart after exhausting attempts - bound plugin startup with an explicit 30s timeout - reject backslash paths and cap decompressed bytes by actual size in pkgtar; only mark a package Signed when a verifier actually ran and passed - delete plugin_state on uninstall; validate webhook keys against the manifest and cap webhook body size; restart an active plugin when a new version installs Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> * fix(plugins): sanitize plugin-id path sinks for CodeQL Pin the plugin id to a single path segment with filepath.Base at recordPath/ configPath and pkgtar.Remove — a barrier the taint analysis recognizes. Behaviour-identical for valid ids (already validated by safePluginID); clears the go/path-injection alerts. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> * fix(plugins): inline plugin-id validation at path sinks for CodeQL filepath.Base does not strip '..', so it is not a path-injection barrier. Call the recognized safePluginID / id-pattern guard inline in Delete/GetConfig/ SetConfig and pkgtar.Remove — the same barrier that already cleared the Get sink. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> * fix(plugins): address code-review findings (backend) - add capability-gating tests for the Host RPC service (was untested) - close a stop-during-start orphan race so Disable/Uninstall can't leave a supervised subprocess behind - on failed upgrade, remove only the new version dir (not all versions + data); restart the prior version - validate manifest version and reject ids ending in .config (record-file alias) - make Deliverer.Refresh non-blocking and cancel in-flight delivery on worker stop - clamp out-of-range webhook status, skip-and-warn corrupt records in List, write records atomically, semver-aware sideload version pick Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> * fix(plugins): address code-review findings (frontend) - wrap plugin routes + settings routes in a shared PluginErrorBoundary so a throwing plugin page shows a fallback instead of white-screening the SPA - fix disable/re-enable in one session: keep the cached registration and re-run initialize instead of relying on a cached ES-module re-import - prefix root-relative bundle/style URLs with apiBaseUrl (split-origin/Tauri) - surface partial-install warnings via toast.warning instead of a bare success - dedupe/remove injected style links on unload; gate PluginNavItems on the flag Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> * fix(plugins): serialize per-plugin lifecycle + guard config writes - add a per-plugin keyed lock in Service so concurrent Enable/Disable/Install/ Uninstall/UpdateConfig for one id are serialized (no status-machine race) - take s.mu in SetConfig/GetConfig and write config atomically - bound each plugin initialize() with a timeout so one hung bundle can't stall the rest of the boot load Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> * test(plugins): split service_test.go to satisfy file-length lint Move the install/URL test group into service_install_test.go; no logic change. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> * ci: build and ship the e2e fixture plugin package to E2E shards The plugin E2E specs' global-setup requires apps/backend/.build/kandev-plugin-e2e-1.0.0.tar.gz. Build it in the E2E build job, upload it as an artifact, and download it in both the regular and container shard jobs. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> --------- Co-authored-by: Kandev Agent <agent@kandev.dev> Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
295 lines
9.1 KiB
Go
295 lines
9.1 KiB
Go
package store
|
|
|
|
import (
|
|
"fmt"
|
|
"os"
|
|
"path/filepath"
|
|
"regexp"
|
|
"strings"
|
|
"sync"
|
|
|
|
"go.uber.org/zap"
|
|
"gopkg.in/yaml.v3"
|
|
|
|
"github.com/kandev/kandev/internal/common/logger"
|
|
)
|
|
|
|
// idPattern mirrors internal/plugins/manifest's plugin id pattern:
|
|
// lowercase alphanumerics, dots, underscores, and hyphens, starting with a
|
|
// lowercase alphanumeric. manifest.Validate already enforces this shape
|
|
// upstream during install; this is the sink-level guard every FSStore
|
|
// method re-applies to id before building a filesystem path from it, so a
|
|
// malformed id (e.g. containing "..", "/", or "\") can never reach disk
|
|
// I/O regardless of how it got here.
|
|
var idPattern = regexp.MustCompile(`^[a-z0-9][a-z0-9._-]*$`)
|
|
|
|
// dotConfigSuffix is the suffix a plugin id must never end with: id+".yml"
|
|
// would then collide with the "<id>.config.yml" naming convention
|
|
// isRecordFile uses to distinguish a plugin's own record from another
|
|
// plugin's operator config, silently hiding the record from List() (see
|
|
// isRecordFile).
|
|
const dotConfigSuffix = ".config"
|
|
|
|
// safePluginID rejects any id that is not a single, clean path segment
|
|
// matching idPattern, or that ends in ".config" (see dotConfigSuffix). It is
|
|
// the guard called by every FSStore method before recordPath/configPath
|
|
// build a path from id.
|
|
func safePluginID(id string) error {
|
|
if !idPattern.MatchString(id) {
|
|
return fmt.Errorf("plugins: invalid plugin id %q", id)
|
|
}
|
|
if strings.HasSuffix(id, dotConfigSuffix) {
|
|
return fmt.Errorf("plugins: invalid plugin id %q: must not end in %q", id, dotConfigSuffix)
|
|
}
|
|
return nil
|
|
}
|
|
|
|
// FSStore persists plugin installation records under dir as "{id}.yml" (the
|
|
// Record) and "{id}.config.yml" (operator-editable config). FSStore
|
|
// implements Store.
|
|
type FSStore struct {
|
|
dir string
|
|
log *logger.Logger
|
|
|
|
// mu guards every disk write (Save, Delete, SetConfig) and read
|
|
// (GetConfig) that touches a plugin's record/config files, so
|
|
// concurrent callers (e.g. two operator UpdateConfig PATCH requests for
|
|
// the same id) can never interleave writes to the same path. Writers
|
|
// take Lock; GetConfig takes RLock since it only reads.
|
|
mu sync.RWMutex
|
|
}
|
|
|
|
// NewFSStore returns a store that persists records under dir.
|
|
func NewFSStore(dir string) *FSStore {
|
|
return &FSStore{dir: dir}
|
|
}
|
|
|
|
// SetLogger wires a logger List uses to warn about a corrupt/unreadable
|
|
// record file it has to skip. Optional: a nil (default, unset) logger makes
|
|
// List silently skip instead of warning.
|
|
func (s *FSStore) SetLogger(log *logger.Logger) {
|
|
s.log = log
|
|
}
|
|
|
|
// recordPath returns the on-disk path for a plugin's installation record.
|
|
// Every public method validates id with safePluginID before reaching here.
|
|
func (s *FSStore) recordPath(id string) string {
|
|
return filepath.Join(s.dir, id+".yml")
|
|
}
|
|
|
|
// configPath returns the on-disk path for a plugin's operator config.
|
|
func (s *FSStore) configPath(id string) string {
|
|
return filepath.Join(s.dir, id+".config.yml")
|
|
}
|
|
|
|
// writeRecord marshals record to YAML and writes it to disk, creating the
|
|
// store directory if needed. Used both to create a fresh record (Install)
|
|
// and to persist updates to an existing one (Save). The write is atomic
|
|
// (tmp file + rename within the same directory) so a crash mid-write can
|
|
// never leave a half-written "<id>.yml" for List()/Get() to trip over.
|
|
func (s *FSStore) writeRecord(record *Record) error {
|
|
if err := safePluginID(record.ID); err != nil {
|
|
return err
|
|
}
|
|
if err := os.MkdirAll(s.dir, 0o755); err != nil {
|
|
return fmt.Errorf("create plugin store dir: %w", err)
|
|
}
|
|
data, err := yaml.Marshal(record)
|
|
if err != nil {
|
|
return fmt.Errorf("marshal plugin record: %w", err)
|
|
}
|
|
if err := writeFileAtomic(s.dir, s.recordPath(record.ID), data, 0o600); err != nil {
|
|
return fmt.Errorf("write plugin record: %w", err)
|
|
}
|
|
return nil
|
|
}
|
|
|
|
// writeFileAtomic writes data to path via a temp file created in dir (the
|
|
// same filesystem/directory as path, so the final os.Rename is atomic) and
|
|
// an fsync-then-rename, so a reader (List/Get) never observes a
|
|
// partially-written file. The temp file is removed on any failure before
|
|
// the rename.
|
|
func writeFileAtomic(dir, path string, data []byte, mode os.FileMode) error {
|
|
tmp, err := os.CreateTemp(dir, ".tmp-*")
|
|
if err != nil {
|
|
return fmt.Errorf("create temp file: %w", err)
|
|
}
|
|
tmpPath := tmp.Name()
|
|
succeeded := false
|
|
defer func() {
|
|
if !succeeded {
|
|
_ = os.Remove(tmpPath)
|
|
}
|
|
}()
|
|
|
|
if _, err := tmp.Write(data); err != nil {
|
|
_ = tmp.Close()
|
|
return fmt.Errorf("write temp file: %w", err)
|
|
}
|
|
if err := tmp.Sync(); err != nil {
|
|
_ = tmp.Close()
|
|
return fmt.Errorf("sync temp file: %w", err)
|
|
}
|
|
if err := tmp.Close(); err != nil {
|
|
return fmt.Errorf("close temp file: %w", err)
|
|
}
|
|
if err := os.Chmod(tmpPath, mode); err != nil {
|
|
return fmt.Errorf("chmod temp file: %w", err)
|
|
}
|
|
if err := os.Rename(tmpPath, path); err != nil {
|
|
return fmt.Errorf("rename temp file: %w", err)
|
|
}
|
|
succeeded = true
|
|
return nil
|
|
}
|
|
|
|
// Get loads the record for id from disk.
|
|
func (s *FSStore) Get(id string) (*Record, error) {
|
|
if err := safePluginID(id); err != nil {
|
|
return nil, err
|
|
}
|
|
data, err := os.ReadFile(s.recordPath(id))
|
|
if err != nil {
|
|
if os.IsNotExist(err) {
|
|
return nil, ErrNotFound
|
|
}
|
|
return nil, fmt.Errorf("read plugin record: %w", err)
|
|
}
|
|
var record Record
|
|
if err := yaml.Unmarshal(data, &record); err != nil {
|
|
return nil, fmt.Errorf("unmarshal plugin record: %w", err)
|
|
}
|
|
return &record, nil
|
|
}
|
|
|
|
// List returns every installed plugin's record. A record file that fails to
|
|
// read or parse (corrupt, stray, or written by a future incompatible
|
|
// version) is skipped and warned about rather than failing List
|
|
// wholesale — the whole plugin subsystem's boot path depends on List()
|
|
// succeeding (Registry.Load), so one bad file must not disable every other
|
|
// installed plugin.
|
|
func (s *FSStore) List() ([]*Record, error) {
|
|
entries, err := os.ReadDir(s.dir)
|
|
if err != nil {
|
|
if os.IsNotExist(err) {
|
|
return nil, nil
|
|
}
|
|
return nil, fmt.Errorf("read plugin store dir: %w", err)
|
|
}
|
|
|
|
var records []*Record
|
|
for _, entry := range entries {
|
|
if entry.IsDir() || !isRecordFile(entry.Name()) {
|
|
continue
|
|
}
|
|
id := strings.TrimSuffix(entry.Name(), ".yml")
|
|
record, err := s.Get(id)
|
|
if err != nil {
|
|
s.warnSkippedRecord(id, err)
|
|
continue
|
|
}
|
|
records = append(records, record)
|
|
}
|
|
return records, nil
|
|
}
|
|
|
|
// warnSkippedRecord logs (if a logger is wired via SetLogger) that List
|
|
// skipped id's record file rather than failing outright.
|
|
func (s *FSStore) warnSkippedRecord(id string, err error) {
|
|
if s.log == nil {
|
|
return
|
|
}
|
|
s.log.Warn("plugins: skipping unreadable plugin record",
|
|
zap.String("plugin_id", id), zap.Error(err))
|
|
}
|
|
|
|
// isRecordFile reports whether name is a plugin record file ("{id}.yml"),
|
|
// excluding config files ("{id}.config.yml").
|
|
func isRecordFile(name string) bool {
|
|
return strings.HasSuffix(name, ".yml") && !strings.HasSuffix(name, ".config.yml")
|
|
}
|
|
|
|
// Save writes record to disk, creating it if it does not already exist or
|
|
// overwriting it (in full) if it does. Used both to persist a fresh install
|
|
// and to persist status/metadata updates to an existing record.
|
|
func (s *FSStore) Save(record *Record) error {
|
|
s.mu.Lock()
|
|
defer s.mu.Unlock()
|
|
return s.writeRecord(record)
|
|
}
|
|
|
|
// Delete removes the record and its operator config for id.
|
|
func (s *FSStore) Delete(id string) error {
|
|
if err := safePluginID(id); err != nil {
|
|
return err
|
|
}
|
|
s.mu.Lock()
|
|
defer s.mu.Unlock()
|
|
|
|
if _, err := s.Get(id); err != nil {
|
|
return err
|
|
}
|
|
|
|
if err := os.Remove(s.recordPath(id)); err != nil {
|
|
return fmt.Errorf("delete plugin record: %w", err)
|
|
}
|
|
if err := os.Remove(s.configPath(id)); err != nil && !os.IsNotExist(err) {
|
|
return fmt.Errorf("delete plugin config: %w", err)
|
|
}
|
|
return nil
|
|
}
|
|
|
|
// GetConfig returns the operator-editable config for id, or an empty map if
|
|
// no config has been set yet.
|
|
func (s *FSStore) GetConfig(id string) (map[string]any, error) {
|
|
if err := safePluginID(id); err != nil {
|
|
return nil, err
|
|
}
|
|
s.mu.RLock()
|
|
defer s.mu.RUnlock()
|
|
|
|
if _, err := s.Get(id); err != nil {
|
|
return nil, err
|
|
}
|
|
data, err := os.ReadFile(s.configPath(id))
|
|
if err != nil {
|
|
if os.IsNotExist(err) {
|
|
return map[string]any{}, nil
|
|
}
|
|
return nil, fmt.Errorf("read plugin config: %w", err)
|
|
}
|
|
var cfg map[string]any
|
|
if err := yaml.Unmarshal(data, &cfg); err != nil {
|
|
return nil, fmt.Errorf("unmarshal plugin config: %w", err)
|
|
}
|
|
if cfg == nil {
|
|
cfg = map[string]any{}
|
|
}
|
|
return cfg, nil
|
|
}
|
|
|
|
// SetConfig replaces the operator-editable config for id. The write is
|
|
// atomic (tmp file + rename within the store dir, via writeFileAtomic), the
|
|
// same guarantee writeRecord gives Save, and is additionally serialized by
|
|
// s.mu so two concurrent SetConfig calls for the same id (e.g. two operator
|
|
// UpdateConfig PATCH requests racing) can never interleave.
|
|
func (s *FSStore) SetConfig(id string, config map[string]any) error {
|
|
if err := safePluginID(id); err != nil {
|
|
return err
|
|
}
|
|
s.mu.Lock()
|
|
defer s.mu.Unlock()
|
|
|
|
if _, err := s.Get(id); err != nil {
|
|
return err
|
|
}
|
|
data, err := yaml.Marshal(config)
|
|
if err != nil {
|
|
return fmt.Errorf("marshal plugin config: %w", err)
|
|
}
|
|
if err := writeFileAtomic(s.dir, s.configPath(id), data, 0o644); err != nil {
|
|
return fmt.Errorf("write plugin config: %w", err)
|
|
}
|
|
return nil
|
|
}
|