Issues

• Restart fail-fast on a repeated failure signature + expose session Disposition to hosts

  2389-research/tracker#736 · Filter by author clintecker opened on Oct 21, 2026

• Pause at a human gate when there is no terminal; answer with tracker answer (agent-driven Tracker 2a)

  2389-research/tracker#732 · Filter by author clintecker opened on Oct 20, 2026

• build_product verify.sh: an UNCHANGED cargo member crate is tested whole, so a test-less milestone greens on the prior suite (#901 scoping hole)

  2389-research/tracker#726 · Filter by author clintecker opened on Oct 19, 2026

• llm: Gemini 'prepayment credits are depleted' (429 RESOURCE_EXHAUSTED) is classified as a retryable RateLimitError, not billing exhaustion

  bug
  2389-research/tracker#722 · Filter by author clintecker opened on Oct 19, 2026

• engine: a freeform reply that matches no label routes by lexical target order

  2389-research/tracker#721 · Filter by author clintecker opened on Oct 19, 2026

• Subgraph events carry the child engine's run_id, not the parent's (#657 follow-up)

  2389-research/tracker#718 · Filter by author harperreed opened on Oct 18, 2026

• run.json totals count subgraph agents' tokens twice

  2389-research/tracker#717 · Filter by author harperreed opened on Oct 18, 2026

• pipeline: tool_command trusts context values a resumed checkpoint restores

  area/enginebugeffort:MP3: low
  2389-research/tracker#714 · Filter by author harperreed opened on Oct 17, 2026

• Port agent/exec core + handler backends to Windows for a full GOOS=windows build

  area/agentenhancementP3: low
  2389-research/tracker#672 · Filter by author clintecker opened on Oct 6, 2026

• build_product: converge tracker's built-in with the tracker-runner fork — structural, testing and quality overhaul (epic, main since v0.73.2)

  area/engineeffort:XLenhancementjiffy:runP1: high
  2389-research/tracker#655 · Filter by author clintecker opened on Sep 18, 2026 · 1 · Trust & verification
  5

• engine: workflow-level progress narrator — a supervisor's running log of where the run is in the overall job (between-node hook, graph-aware)

  area/engineeffort:LenhancementP2: medium
  2389-research/tracker#641 · Filter by author clintecker opened on Sep 17, 2026 · 2 · .dip language & engine expressiveness
  1

• site: local hugo server fails on Hugo 0.166 — text/html content not whitelisted (CI pins 0.161.1)

  bugdocumentationeffort:SP3: low
  2389-research/tracker#638 · Filter by author harperreed opened on Sep 13, 2026 · 4 · Product & 1.0 readiness
  1

• Human-gate artifacts: ship as separate named attachments + emit gate status tokens + structured options (enables tracker-runner #750 gate redesign)

  area/enginearea/tuieffort:MenhancementP2: medium
  2389-research/tracker#634 · Filter by author clintecker opened on Sep 5, 2026 · Transport boundary
  1

• Meta Muse models: unpriced (budget guard reads $0) and no provider adapter — reachability + the unpriced-model-under-a-cost-ceiling hole

  area/adapterarea/agenteffort:MenhancementP2: mediumupstream
  2389-research/tracker#627 · Filter by author clintecker opened on Sep 2, 2026 · 1 · Trust & verification
  1

• Epic: Adversarial Review — evidence-grounded typed-verdict code review

  effort:LenhancementP2: medium
  2389-research/tracker#625 · Filter by author clintecker opened on Aug 26, 2026 · 1 · Trust & verification
  1

• build_product: adopt adversarial-review (typed verdicts + deterministic FP gate) in the review→fix phase

  effort:MenhancementP2: medium
  2389-research/tracker#624 · Filter by author clintecker opened on Aug 26, 2026 · 1 · Trust & verification
  1

• product: publish the first SWE-bench score — site Benchmarks slot + reframe swebench as evidence (gated on #465)

  documentationeffort:MenhancementP3: lowphase:2
  2389-research/tracker#589 · Filter by author clintecker opened on Aug 21, 2026 · 4 · Product & 1.0 readiness

• decision(naming): qualify-forever vs brand≠command (Option C) vs full rename — evaluate the gate

  effort:Sjiffy:runP2: mediumphase:1question
  2389-research/tracker#586 · Filter by author clintecker opened on Aug 21, 2026 · 4 · Product & 1.0 readiness

• product(naming): recon 3–5 coined name candidates + availability checks (feeds the rename gate)

  effort:MenhancementP2: mediumphase:1
  2389-research/tracker#585 · Filter by author clintecker opened on Aug 21, 2026 · 4 · Product & 1.0 readiness

• product(naming): stand up tracker.2389.ai → docs redirect + CNAME

  documentationeffort:SP2: mediumphase:0
  2389-research/tracker#583 · Filter by author clintecker opened on Aug 21, 2026 · 4 · Product & 1.0 readiness

• product(site): build a /trust page linking the threat-model docs

  documentationeffort:MP2: mediumphase:1
  2389-research/tracker#581 · Filter by author clintecker opened on Aug 21, 2026 · 4 · Product & 1.0 readiness

• Verification cost ~= implementation cost per milestone

  area/engineeffort:MenhancementP2: mediumperformance
  2389-research/tracker#490 · Filter by author clintecker opened on Jul 21, 2026 · 1 · Trust & verification
  4

• benchmarks: first scored SWE-bench Verified run

  effort:LenhancementP2: mediumtesting
  2389-research/tracker#465 · Filter by author harperreed opened on Jul 6, 2026 · 4 · Product & 1.0 readiness
  3

• product: tracker-swebench + tracker-conformance in-repo read as research tooling — relocate or reframe

  choreP3: lowquestion
  2389-research/tracker#464 · Filter by author harperreed opened on Jul 6, 2026 · 4 · Product & 1.0 readiness
  2

• cli: run-flag surface (~36 flags) needs presets or progressive disclosure (companion to #396)

  area/clieffort:MenhancementP2: medium
  2389-research/tracker#463 · Filter by author harperreed opened on Jul 6, 2026 · 3 · Library, CLI & refactor quality
  1