Add wisp-gc doctor check, integrate into gt doctor --fix (gt-psj76.2)

- Implement WispGCCheck in internal/doctor/wisp_check.go
  - Scans rigs for wisps older than 1 hour threshold
  - Fix runs `bd --no-daemon wisp gc` in each affected rig
- Register wisp-gc check in gt doctor
- Update help text to document cleanup checks
- Simplify Deacon patrol session-gc step to just use gt doctor --fix

Now `gt doctor --fix` handles all cleanup:
- orphan-sessions: Kill orphaned tmux sessions
- orphan-processes: Kill orphaned Claude processes
- wisp-gc: Garbage collect abandoned wisps (>1h)

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

.beads/formulas/mol-deacon-patrol.formula.json

@@ -36,7 +36,7 @@
       "id": "session-gc",
       "title": "Clean dead sessions",
       "needs": ["orphan-check"],
-      "description": "Clean dead sessions and orphaned state.\n\nGarbage collect terminated sessions and orphaned artifacts:\n\n## 1. Clean orphaned sessions and processes\n\n```bash\n# Check for orphans (dry run)\ngt doctor -v\n\n# Fix orphaned sessions and processes\ngt doctor --fix\n```\n\nThis handles:\n- Orphaned tmux sessions (gt-* not matching valid rig/role patterns)\n- Orphaned Claude processes (no tmux parent)\n\n## 2. Clean orphaned wisps\n\n```bash\n# Clean wisps older than 1 hour (default threshold)\nbd --no-daemon wisp gc\n\n# Or with custom age threshold\nbd --no-daemon wisp gc --age 2h\n\n# Preview what would be cleaned\nbd --no-daemon wisp gc --dry-run\n```\n\nThe `--no-daemon` flag is required when the bd daemon is running, since wisp gc needs direct database access.\n\n## Notes\n\n- Preserve audit trail - only clean confirmed dead sessions\n- Old logs and temp files: not yet implemented\n- Completed molecule archival: handled by bd mol squash/burn"
+      "description": "Clean dead sessions and orphaned state.\n\nRun `gt doctor --fix` to handle all cleanup:\n\n```bash\n# Preview what needs cleaning\ngt doctor -v\n\n# Fix everything\ngt doctor --fix\n```\n\nThis handles:\n- **orphan-sessions**: Kill orphaned tmux sessions (gt-* not matching valid patterns)\n- **orphan-processes**: Kill orphaned Claude processes (no tmux parent)\n- **wisp-gc**: Garbage collect abandoned wisps (>1h old)\n\nAll cleanup is handled by doctor checks - no need to run separate commands."
     },
     {
       "id": "context-check",

internal/cmd/doctor.go

@@ -24,6 +24,11 @@ var doctorCmd = &cobra.Command{
 Doctor checks for common configuration issues, missing files,
 and other problems that could affect workspace operation.
 
+Cleanup checks (fixable):
+  - orphan-sessions          Detect orphaned tmux sessions
+  - orphan-processes         Detect orphaned Claude processes
+  - wisp-gc                  Detect and clean abandoned wisps (>1h)
+
 Patrol checks:
   - patrol-molecules-exist   Verify patrol molecules exist
   - patrol-hooks-wired       Verify daemon triggers patrols
@@ -66,6 +71,7 @@ func runDoctor(cmd *cobra.Command, args []string) error {
 	d.Register(doctor.NewBeadsDatabaseCheck())
 	d.Register(doctor.NewOrphanSessionCheck())
 	d.Register(doctor.NewOrphanProcessCheck())
+	d.Register(doctor.NewWispGCCheck())
 	d.Register(doctor.NewBranchCheck())
 	d.Register(doctor.NewBeadsSyncOrphanCheck())
 	d.Register(doctor.NewIdentityCollisionCheck())

internal/doctor/wisp_check.go

@@ -1,14 +1,143 @@
 package doctor
 
-// Legacy wisp directory checks removed.
+import (
-// Wisps are now just a flag on regular beads issues (Wisp: true).
+	"bufio"
-// Hook files are stored in .beads/ alongside other beads data.
+	"encoding/json"
-//
+	"fmt"
-// These checks were for the old .beads-wisp/ directory infrastructure:
+	"os"
-// - WispExistsCheck: checked if .beads-wisp/ exists
+	"os/exec"
-// - WispGitCheck: checked if .beads-wisp/ is a git repo
+	"path/filepath"
-// - WispOrphansCheck: checked for old wisps
+	"time"
-// - WispSizeCheck: checked size of .beads-wisp/
+)
-// - WispStaleCheck: checked for inactive wisps
+
-//
+// WispGCCheck detects and cleans orphaned wisps that are older than a threshold.
-// All removed as of the wisp simplification (gt-5klh, bd-bkul).
+// Wisps are ephemeral issues (Wisp: true flag) used for patrol cycles and
+// operational workflows that shouldn't accumulate.
+type WispGCCheck struct {
+	FixableCheck
+	threshold     time.Duration
+	abandonedRigs map[string]int // rig -> count of abandoned wisps
+}
+
+// NewWispGCCheck creates a new wisp GC check with 1 hour threshold.
+func NewWispGCCheck() *WispGCCheck {
+	return &WispGCCheck{
+		FixableCheck: FixableCheck{
+			BaseCheck: BaseCheck{
+				CheckName:        "wisp-gc",
+				CheckDescription: "Detect and clean orphaned wisps (>1h old)",
+			},
+		},
+		threshold:     1 * time.Hour,
+		abandonedRigs: make(map[string]int),
+	}
+}
+
+// Run checks for abandoned wisps in each rig.
+func (c *WispGCCheck) Run(ctx *CheckContext) *CheckResult {
+	c.abandonedRigs = make(map[string]int)
+
+	rigs, err := discoverRigs(ctx.TownRoot)
+	if err != nil {
+		return &CheckResult{
+			Name:    c.Name(),
+			Status:  StatusError,
+			Message: "Failed to discover rigs",
+			Details: []string{err.Error()},
+		}
+	}
+
+	if len(rigs) == 0 {
+		return &CheckResult{
+			Name:    c.Name(),
+			Status:  StatusOK,
+			Message: "No rigs configured",
+		}
+	}
+
+	var details []string
+	totalAbandoned := 0
+
+	for _, rigName := range rigs {
+		rigPath := filepath.Join(ctx.TownRoot, rigName)
+		count := c.countAbandonedWisps(rigPath)
+		if count > 0 {
+			c.abandonedRigs[rigName] = count
+			totalAbandoned += count
+			details = append(details, fmt.Sprintf("%s: %d abandoned wisp(s)", rigName, count))
+		}
+	}
+
+	if totalAbandoned > 0 {
+		return &CheckResult{
+			Name:    c.Name(),
+			Status:  StatusWarning,
+			Message: fmt.Sprintf("%d abandoned wisp(s) found (>1h old)", totalAbandoned),
+			Details: details,
+			FixHint: "Run 'gt doctor --fix' to garbage collect orphaned wisps",
+		}
+	}
+
+	return &CheckResult{
+		Name:    c.Name(),
+		Status:  StatusOK,
+		Message: "No abandoned wisps found",
+	}
+}
+
+// countAbandonedWisps counts wisps older than the threshold in a rig.
+func (c *WispGCCheck) countAbandonedWisps(rigPath string) int {
+	// Check the beads database for wisps
+	issuesPath := filepath.Join(rigPath, ".beads", "issues.jsonl")
+	file, err := os.Open(issuesPath)
+	if err != nil {
+		return 0 // No issues file
+	}
+	defer file.Close()
+
+	cutoff := time.Now().Add(-c.threshold)
+	count := 0
+
+	scanner := bufio.NewScanner(file)
+	for scanner.Scan() {
+		line := scanner.Text()
+		if line == "" {
+			continue
+		}
+
+		var issue struct {
+			ID        string    `json:"id"`
+			Status    string    `json:"status"`
+			Wisp      bool      `json:"wisp"`
+			UpdatedAt time.Time `json:"updated_at"`
+		}
+		if err := json.Unmarshal([]byte(line), &issue); err != nil {
+			continue
+		}
+
+		// Count wisps that are not closed and older than threshold
+		if issue.Wisp && issue.Status != "closed" && !issue.UpdatedAt.IsZero() && issue.UpdatedAt.Before(cutoff) {
+			count++
+		}
+	}
+
+	return count
+}
+
+// Fix runs bd wisp gc in each rig with abandoned wisps.
+func (c *WispGCCheck) Fix(ctx *CheckContext) error {
+	var lastErr error
+
+	for rigName := range c.abandonedRigs {
+		rigPath := filepath.Join(ctx.TownRoot, rigName)
+
+		// Run bd --no-daemon wisp gc
+		cmd := exec.Command("bd", "--no-daemon", "wisp", "gc")
+		cmd.Dir = rigPath
+		if output, err := cmd.CombinedOutput(); err != nil {
+			lastErr = fmt.Errorf("%s: %v (%s)", rigName, err, string(output))
+		}
+	}
+
+	return lastErr
+}