fix: honor --incognito by disabling the experience tracker (plan 009)

[DavertMik]

Implements plan 009 (plans/009-wire-incognito-flag.md) — a P2 bug fix (deferred finding CORRECTNESS-03).

Problem

explorbot --incognito promises "Run without recording experiences". The flag travels CLI → ExplorBot → Explorer, and Explorer calls new StateManager({ incognito: ... }) — but StateManager's constructor took no parameters, so the object argument was silently ignored and the ExperienceTracker it created was fully enabled. Incognito runs read and wrote ./experience/ exactly like normal runs; the user's explicit "don't learn from this session" request was dropped without any error (no typecheck gate to catch it).

Fix

Thread incognito into the constructor and pass it through to the tracker, which already supports { disabled }:

constructor(options: { incognito?: boolean } = {}) {
  this.experienceTracker = new ExperienceTracker({ disabled: options.incognito });

disabled: undefined collapses to false via the tracker's options.disabled ?? false, so the two argumentless new StateManager() call sites keep today's behavior. explorer.ts:85 was already correct and needs no edit.

Verification

• New tests: new StateManager({ incognito: true }) → disabled tracker; default → enabled. (Asserted via the plan's sanctioned (tracker as any).disabled fallback.)
• bun test tests/unit → 728 pass, 0 fail; bun run lint clean.

Known gap (out of scope, per plan)

Agents that construct their own ExperienceTracker (Navigator, Captain) bypass this flag until the DI-consolidation plan lands.

🤖 Generated with Claude Code

[github-actions]

Explorbot Self-Regression

Commit 5050c6c · run

Scenario	Result	Attempts	Duration
basic (native)	FAIL	3/3	26m

Attempt details

• basic (native) attempt 1 — FAIL: login evidence: PASS (post-login plan=true, post-login research=true); research: PASS (files=8, wellFormed=true, keywords=7/3); scenarios: PASS (tests=7/5, features=4/3); tests passed: FAIL (3 passed, 2 failed (reporter: 3 passed, 2 failed))
• basic (native) attempt 2 — FAIL: login evidence: PASS (post-login plan=true, post-login research=true); research: PASS (files=7, wellFormed=true, keywords=7/3); scenarios: PASS (tests=7/5, features=4/3); tests passed: FAIL (3 passed, 2 failed (reporter: 3 passed, 2 failed))
• basic (native) attempt 3 — FAIL: login evidence: PASS (post-login plan=true, post-login research=true); research: PASS (files=6, wellFormed=true, keywords=7/3); scenarios: PASS (tests=6/5, features=4/3); tests passed: FAIL (4 passed, 1 failed (reporter: 4 passed, 1 failed))

---

Session analysis — basic (native):

Session Analysis

The Issues list page core flows work: creating new issues, searching by text, filtering by label, and viewing issue details all function correctly. The status filter functionality works (ET-3 verification passed), but the automation stalled after verification, requiring investigation.

Coverage

• Pages: /issues, /issues?new=1, Fix Bun compatibility: add ora-classic and update @testomatio/reporter #4
• Features: Create issue form, text search, label filter, status filter, issue detail view

What works

• New Issue form — ET-1 Create a new issue via toolbar button
• Text search — ET-2 Search for issues by text and verify filter works
• Label filter — ET-4 Filter issues by label and verify results
• Issue detail view — ET-5 Open an issue from the list to view its details

Execution Issues

• ET-3 Filter issues by status and verify results — verification passed (filter worked, correct issues shown) but automation stalled with "No browser progress after repeated attempts" after the test assertions succeeded

[DavertMik]

Closing — re-creating plans 009 and 010 as separate independent PRs off main (this branch had both entangled).