- Python 3.12+
- Agent Orchestrator (
ao) CLI installed and authenticated - Claude Code (
claude) CLI installed ghCLI authenticated with access tojleechanorg/mctrl_test- System loadavg < 20 (check with
sysctl -n vm.loadavg)
- Test Name: level_up_entry_offer_pr6958
- Run ID:
level_up_entry_offer_pr6958-014-20260520T201714 - Iteration: 14
- Bundle Version: 1.2.0
- Collected At (UTC): 2026-05-20T20:17:14.029698+00:00
- Repository: worldarchitect.ai
- Branch: fix/6926-review-comments
WorldArchitect PR 6958 evidence reproduction
PR: https://github.com/jleechanorg/worldarchitect.ai/pull/6958
Current PR HEAD: db5698c7af88592f20624dd69fb9047a1b65bca7
Evidence execution SHA: db5698c7af88592f20624dd69fb9047a1b65bca7
Evidence is fresh — no staleness gap.
PR: https://github.com/jleechanorg/worldarchitect.ai/pull/6958 Commit: c81ee83e0271f37b2bac52d7c5a2df3a4b334d29 Branch: fix/6926-review-comments Date: 2026-05-20
PR: https://github.com/jleechanorg/worldarchitect.ai/pull/6963 Repository: jleechanorg/worldarchitect.ai Worktree: /private/tmp/worldarchitect-skills-pr Collected: 2026-05-20 UTC
PR 6963 is an agent-facing docs/configuration change set. It does not change production runtime code under mvp_site/**; the only mvp_site/ file changed is mvp_site/CLAUDE.md, a documentation/policy file. No UI/user-visible runtime flow changes are claimed, so no mp4/gif/cast evidence is required for this docs/config evidence claim.
PR: https://github.com/jleechanorg/worldarchitect.ai/pull/6960 Evidence commit: d9a8f3e4a29d52c986be0b910e51aa469af8193f Branch: fix/pr6906-successor-cc-guards
git clone https://github.com/jleechanorg/worldarchitect.ai.git- Repository: worldarchitect.ai
- Branch: fix/6926-review-comments
- PR: https://github.com/jleechanorg/worldarchitect.ai/pull/6958
- Collected: 2026-05-19T18:18:07Z
- HEAD: 0ef0cc1a8a
| { | |
| "benchmark": "test_character_creation_three_flows 3-run cache timing", | |
| "git_head": "6ab6563f86c5b609d9299c34f6e48d45f5997ed9", | |
| "branch": "feat/llm-response-cache-2026-05-16", | |
| "description": "3 consecutive runs of testing_mcp/core/test_character_creation_three_flows.py with WORLDAI_TEST_CACHE=read_write. Cold = cache cleared, Warm = cache from cold, Hot = cache from warm. Each run creates new campaigns with unique IDs, so cross-run cache hits rely on prompt fingerprint matching (same user_input + model + mode) not campaign_id matching.", | |
| "runs": [ | |
| { | |
| "run": "cold", | |
| "elapsed_seconds": 484.0, | |
| "cache_entries_after": 34, |
| { | |
| "benchmark": "test_character_creation_three_flows 3-run cache timing", | |
| "git_head": "6ab6563f86c5b609d9299c34f6e48d45f5997ed9", | |
| "branch": "feat/llm-response-cache-2026-05-16", | |
| "runs": [ | |
| { | |
| "run": "cold", | |
| "elapsed_seconds": 484.0, | |
| "cache_entries_after": 34, | |
| "exit_code": 1, |