Issues

• [Loop Eval][Replay] Mine loop-control bug candidates with repeated prompt replay

  area:evaluation-harnessarea:loop-controlkind:evaluationkind:testingpriority:p1-ga
  agentic-in/inferoa#54 · Filter by author Xunzhuo opened on Jun 13, 2026 · GA-M3 SWE-bench + Internal Corpus
  2

• [Loop Eval][Discover] Run tokenmaxxing/context optimization research loop

  area:evaluation-harnessarea:loop-controlarea:tokenmaxxing-srkind:evaluationkind:researchpriority:p1-ga
  agentic-in/inferoa#53 · Filter by author Xunzhuo opened on Jun 13, 2026 · GA-M3 SWE-bench + Internal Corpus
  1

• [Loop Eval][Deliver] Close a real loop-control bug/use case with verification

  area:evaluation-harnessarea:loop-controlkind:evaluationkind:testingpriority:p1-ga
  agentic-in/inferoa#52 · Filter by author Xunzhuo opened on Jun 13, 2026 · GA-M3 SWE-bench + Internal Corpus
  1

• [GA-M4][evaluation] Compare final acceptance against Terminal, SWE, and internal evidence

  area:evaluation-harnessarea:release-opskind:evaluationpriority:p1-ga
  agentic-in/inferoa#51 · Filter by author Xunzhuo opened on Jun 12, 2026 · GA-M4 Acceptance & Release

• [GA-M4][testing] Add evidence index consistency checks

  area:evaluation-harnessarea:release-opskind:testingpriority:p1-ga
  agentic-in/inferoa#50 · Filter by author Xunzhuo opened on Jun 12, 2026 · GA-M4 Acceptance & Release

• [GA-M4][engineering] Generate GA evidence index from ga-eval artifacts

  area:evaluation-harnessarea:release-opskind:engineeringpriority:p1-ga
  agentic-in/inferoa#49 · Filter by author Xunzhuo opened on Jun 12, 2026 · GA-M4 Acceptance & Release

• [GA-M3][testing] Add internal corpus smoke tasks for verifier, compression, and self-improve

  area:evaluation-harnessarea:self-improvekind:testingpriority:p1-ga
  agentic-in/inferoa#48 · Filter by author Xunzhuo opened on Jun 12, 2026 · GA-M3 SWE-bench + Internal Corpus

• [GA-M3][engineering] Add maker/checker policy for eval tasks

  area:evaluation-harnessarea:loop-controlkind:engineeringpriority:p1-ga
  agentic-in/inferoa#47 · Filter by author Xunzhuo opened on Jun 12, 2026 · GA-M3 SWE-bench + Internal Corpus

• [GA-M3][engineering] Run GA eval tasks in managed worktrees

  area:evaluation-harnessarea:loop-controlkind:engineeringpriority:p1-ga
  agentic-in/inferoa#46 · Filter by author Xunzhuo opened on Jun 12, 2026 · GA-M3 SWE-bench + Internal Corpus

• [GA-M3][evaluation] Implement internal corpus runner

  area:evaluation-harnessarea:self-improvekind:evaluationpriority:p1-ga
  agentic-in/inferoa#45 · Filter by author Xunzhuo opened on Jun 12, 2026 · GA-M3 SWE-bench + Internal Corpus

• [GA-M3][research] Define internal real-loop corpus taxonomy and manifest

  area:evaluation-harnessarea:self-improvekind:researchpriority:p1-ga
  agentic-in/inferoa#44 · Filter by author Xunzhuo opened on Jun 12, 2026 · GA-M3 SWE-bench + Internal Corpus

• [GA-M2][testing] Add offline route replay fixtures for cost and quality guardrails

  area:tokenmaxxing-srkind:testingpriority:p0-blocker
  agentic-in/inferoa#43 · Filter by author Xunzhuo opened on Jun 12, 2026 · GA-M2 Tokenmaxxing + SR Policy

• [GA-M2][evaluation] Generate SR route policy replay report

  area:tokenmaxxing-srkind:evaluationpriority:p0-blocker
  agentic-in/inferoa#42 · Filter by author Xunzhuo opened on Jun 12, 2026 · GA-M2 Tokenmaxxing + SR Policy

• [GA-M2][engineering] Capture SR selected route and model evidence from responses

  area:tokenmaxxing-srkind:engineeringpriority:p0-blocker
  agentic-in/inferoa#41 · Filter by author Xunzhuo opened on Jun 12, 2026 · GA-M2 Tokenmaxxing + SR Policy
  1

• [GA-M2][engineering] Record SR route feature vectors per model request

  area:tokenmaxxing-srkind:engineeringpriority:p0-blocker
  agentic-in/inferoa#40 · Filter by author Xunzhuo opened on Jun 12, 2026 · GA-M2 Tokenmaxxing + SR Policy

• [GA-M1][testing] Add adapter failure-mode fixtures for setup, verifier, and timeout

  area:evaluation-harnesskind:testingpriority:p1-ga
  agentic-in/inferoa#39 · Filter by author Xunzhuo opened on Jun 12, 2026 · GA-M1 Terminal-Bench Loop Eval

• [GA-M1][engineering] Materialize Terminal-Bench task workspaces in managed worktrees

  area:evaluation-harnessarea:loop-controlkind:engineeringpriority:p1-ga
  agentic-in/inferoa#38 · Filter by author Xunzhuo opened on Jun 12, 2026 · GA-M1 Terminal-Bench Loop Eval

• [GA-M1][evaluation] Add baseline runner for Terminal-Bench comparison

  area:evaluation-harnesskind:evaluationpriority:p0-blocker
  agentic-in/inferoa#37 · Filter by author Xunzhuo opened on Jun 12, 2026 · GA-M1 Terminal-Bench Loop Eval

• [GA-M0][engineering] Add GA artifact directory layout and manifest index

  area:evaluation-harnesskind:engineeringpriority:p0-blocker
  agentic-in/inferoa#36 · Filter by author Xunzhuo opened on Jun 12, 2026 · GA-M0 First Closed Loop Evidence Slice

• [GA-M0][testing] Add ga-eval validator fixtures for pass, fail, blocked, and error

  area:evaluation-harnesskind:testingpriority:p0-blocker
  agentic-in/inferoa#35 · Filter by author Xunzhuo opened on Jun 12, 2026 · GA-M0 First Closed Loop Evidence Slice

• [GA-M0][evaluation] Add first internal real-loop closed evidence fixture

  area:evaluation-harnessarea:loop-controlkind:evaluationpriority:p0-blocker
  agentic-in/inferoa#34 · Filter by author Xunzhuo opened on Jun 12, 2026 · GA-M0 First Closed Loop Evidence Slice

• [GA-M0][engineering] Project GoalLoopView and loop metrics into ga-eval/v1

  area:evaluation-harnessarea:loop-controlkind:engineeringpriority:p0-blocker
  agentic-in/inferoa#33 · Filter by author Xunzhuo opened on Jun 12, 2026 · GA-M0 First Closed Loop Evidence Slice

• [GA-M4][engineering] Prepare GA release docs from evidence report

  area:release-opskind:engineeringpriority:p2-followup
  agentic-in/inferoa#32 · Filter by author Xunzhuo opened on Jun 12, 2026 · GA-M4 Acceptance & Release

• [GA-M4][testing] Add GA release gate command

  area:release-opskind:testingpriority:p1-ga
  agentic-in/inferoa#31 · Filter by author Xunzhuo opened on Jun 12, 2026 · GA-M4 Acceptance & Release

• [GA-M4][evaluation] Generate final acceptance evidence report from ga-eval artifacts

  area:release-opsarea:tokenmaxxing-srkind:evaluationpriority:p1-ga
  agentic-in/inferoa#30 · Filter by author Xunzhuo opened on Jun 12, 2026 · GA-M4 Acceptance & Release