Commit Graph

  • d271ef9ecd fix: resume review thread resolution Trevin Chow 2026-08-19 19:46:28 -07:00
  • 3aea3815e9 Address PR review feedback (#1491) Trevin Chow 2026-08-19 17:45:38 -07:00
  • 74dabd38a9 fix(ce-resolve-pr-feedback): keep replies out of pending reviews Trevin Chow 2026-08-19 17:30:38 -07:00
  • 15ab6f76bb fix(skills): complete below-cap refactor sweep (#1490) Trevin Chow 2026-08-19 19:34:17 -07:00
  • 182784978f fix(skills): close review evidence gaps (#1490) tmchow/audit-refactored-skills Trevin Chow 2026-08-19 19:13:02 -07:00
  • ccc6d3407a fix(skills): refresh frame evidence snapshots (#1490) Trevin Chow 2026-08-19 18:13:02 -07:00
  • cbc145e291 fix(skills): derive runtime decisions from resolved facts Trevin Chow 2026-08-19 17:58:19 -07:00
  • b2c2bad66e fix: replace analyzer raw snapshots Trevin Chow 2026-08-19 17:45:06 -07:00
  • 5ad0c23897 fix: normalize Riffrec capture inputs Trevin Chow 2026-08-19 17:32:33 -07:00
  • ae6b014f49 fix: verify polish server URLs Trevin Chow 2026-08-19 17:18:27 -07:00
  • a19cd14d6f fix: align setup browser consumers Trevin Chow 2026-08-19 17:07:45 -07:00
  • 733c9c5cc5 fix: preserve reused polish servers Trevin Chow 2026-08-19 16:57:44 -07:00
  • 5689204e42 fix: resolve follow-up skill feedback Trevin Chow 2026-08-19 16:45:40 -07:00
  • 6e0f46aa8b Address PR review feedback (#1490) Trevin Chow 2026-08-19 16:36:34 -07:00
  • fdd4632329 fix(skills): complete below-cap refactor sweep Trevin Chow 2026-08-19 16:23:50 -07:00
  • ebef4813ab fix(ce-brainstorm): ask only decisions the environment cannot settle (#1487) Trevin Chow 2026-08-19 12:15:16 -07:00
  • e85a7c962b fix(skill-eval): grade lookup-not-ask on the looked-up fact (#1487) tmchow/research-grill-me-vs Trevin Chow 2026-08-19 12:03:10 -07:00
  • ce2b99ff22 fix(ce-brainstorm): ask only decisions the environment cannot settle Trevin Chow 2026-08-19 11:48:11 -07:00
  • 6d2e6b41dd fix(skills): make model-invoked descriptions context pointers (#1486) Trevin Chow 2026-08-19 09:42:09 -07:00
  • d980eaadd4 fix(skills): keep non-derivable activation tokens in descriptions tmchow/skill-description-context-pointers Trevin Chow 2026-08-19 09:37:18 -07:00
  • 0d0fedf769 fix(ce-pov): restore semantic panel triggers and the oracle token Trevin Chow 2026-08-19 09:24:19 -07:00
  • d181ae0eec feat(skill-eval): host-CLI catalog for skill-behavior A/B (#1484) Trevin Chow 2026-08-19 09:17:16 -07:00
  • 12f692ec1f fix(skills): make model-invoked descriptions context pointers Trevin Chow 2026-08-19 09:10:42 -07:00
  • 801659fdb9 fix(skill-eval): the handoff stop cell must detect a committed continuation tmchow/eval-harness-tool-simulation Trevin Chow 2026-08-19 02:14:00 -07:00
  • 634e58bc5c fix(skill-eval): inspect every run commit, and never leave a host running Trevin Chow 2026-08-19 02:08:18 -07:00
  • 3a9ab6d810 fix(skill-eval): fail the pack on a failed arm, and reap the host process tree Trevin Chow 2026-08-19 02:01:10 -07:00
  • 896970d7ed fix(skill-eval): commit grades key on the seed revision, and reads match full paths Trevin Chow 2026-08-19 01:53:52 -07:00
  • 31a547d4cd fix(skill-eval): grade every commit in the run, and four narrower cell defects Trevin Chow 2026-08-19 01:46:33 -07:00
  • 276acbe981 fix(skill-eval): grade the artifact, not the self-report, on three mutation cells Trevin Chow 2026-08-19 01:38:59 -07:00
  • 7f9b06518d refactor(skill-eval): collapse the trailer-presence bookkeeping to direct checks Trevin Chow 2026-08-19 01:38:59 -07:00
  • eb674caca7 fix(skill-eval): pin the two positive probes to something the prompt cannot supply Trevin Chow 2026-08-19 01:30:52 -07:00
  • e75a3a5c00 fix(skill-eval): give read-only restraint cells something that can fail Trevin Chow 2026-08-19 01:22:01 -07:00
  • be77a5ccd0 fix(skill-eval): grade the working tree, and keep PATH shims out of the workspace Trevin Chow 2026-08-19 01:21:50 -07:00
  • 614c726b6c fix(skill-eval): make --read-only actually enforce the fake boundary Trevin Chow 2026-08-19 01:21:50 -07:00
  • acb9f6f3b0 fix(skill-eval): grade the answer on stdout, and require the trailers a grade reads Trevin Chow 2026-08-19 01:21:50 -07:00
  • d2aba7258b fix(skill-eval): grade the host copy, HEAD post arm, and failed cells Trevin Chow 2026-08-19 00:36:18 -07:00
  • c012dbf0bd feat(skill-eval): host-CLI catalog for skill-behavior A/B Trevin Chow 2026-08-19 00:20:59 -07:00
  • 67cc7dc7a1 docs(skill-design): third-sweep learnings — place blocks by executing step, audit pins, size evals to reach (#1483) Trevin Chow 2026-08-18 23:44:38 -07:00
  • bff6a76ee1 fix(skill-design): classify the cap by shipping path, and pins by the failure they record tmchow/playbook-wave4-learnings Trevin Chow 2026-08-18 23:31:18 -07:00
  • bf30692e8f docs(skill-design): record wave-4 learnings — placement by executing step, pins as decisions, eval breadth, the truncation outage Trevin Chow 2026-08-18 23:25:22 -07:00
  • 956087b3e1 fix(lfg): 72% smaller SKILL.md, six required-read references carry the rest (#1479) Trevin Chow 2026-08-18 22:52:01 -07:00
  • 70859c96fb fix(lfg): step 6 gates on anything divergent, not on actionable findings alone tmchow/8kb-lfg Trevin Chow 2026-08-18 22:38:46 -07:00
  • 28d382ad3d Merge origin/main into tmchow/8kb-lfg Trevin Chow 2026-08-18 22:28:05 -07:00
  • ff4d21b8ac fix(ce-brainstorm): 86% smaller SKILL.md, six references carry the phases (#1476) Trevin Chow 2026-08-18 22:10:10 -07:00
  • 312943d4ae fix(ce-plan): audit body pins by provenance, relocate what only wording held in the window (#1475) Trevin Chow 2026-08-18 22:09:51 -07:00
  • 85d44d571d fix(ce-plan): scope the menu return to the copyable-prompt goal arm; make the commented-settings pin falsifiable tmchow/8kb-ce-plan-pins Trevin Chow 2026-08-18 19:19:02 -07:00
  • c5492594fa fix(ce-plan): audit body pins by provenance, relocate what only wording held in the window Trevin Chow 2026-08-18 18:06:24 -07:00
  • 3568cd94a2 fix(ce-plan): 71% smaller SKILL.md, five phase references carry the rest (#1470) Trevin Chow 2026-08-18 22:09:30 -07:00
  • 20f2abb17f Merge origin/main into tmchow/8kb-ce-brainstorm (OVER_BUDGET set) tmchow/8kb-ce-brainstorm Trevin Chow 2026-08-18 22:04:03 -07:00
  • cb230f14eb fix(ce-work): 48% smaller SKILL.md, references carry the execution manual (#1478) Trevin Chow 2026-08-18 22:03:14 -07:00
  • 26f38961ae fix(ce-compound): 90% smaller SKILL.md, references carry the phases (#1477) Trevin Chow 2026-08-18 22:02:58 -07:00
  • ec07a42b97 Merge origin/main into tmchow/8kb-ce-brainstorm (OVER_BUDGET set) Trevin Chow 2026-08-18 21:57:28 -07:00
  • db8bb126a6 Merge origin/main into tmchow/8kb-ce-compound (OVER_BUDGET set) tmchow/8kb-ce-compound Trevin Chow 2026-08-18 21:56:16 -07:00
  • 4df5f99db8 fix(ce-code-review): 86% smaller SKILL.md, four step references carry the rest (#1471) Trevin Chow 2026-08-18 21:54:25 -07:00
  • c389045456 fix(lfg): put every required read before the step that acts on it Trevin Chow 2026-08-18 21:47:07 -07:00
  • ae48e92bfd fix(lfg): 72% smaller SKILL.md, six required-read references carry the rest Trevin Chow 2026-08-18 21:37:49 -07:00
  • 734dd2c815 fix(ce-debug): audit body pins by provenance, relocate what only wording held in the window (#1472) Trevin Chow 2026-08-18 21:10:38 -07:00
  • 2767cc74b8 fix(ce-explain): 32% smaller SKILL.md, the Phase 6 close moves to a required-read reference (#1469) Trevin Chow 2026-08-18 21:09:32 -07:00
  • 3d5897fa56 fix(ce-work): state the strategy-read condition once, for every engine tmchow/8kb-ce-work Trevin Chow 2026-08-18 20:37:58 -07:00
  • 87ecf7cb0f fix(ce-compound): one owner for the Lightweight completion report Trevin Chow 2026-08-18 20:33:02 -07:00
  • 3d86498fd8 fix(ce-work): load the strategy reference before the inline-vs-worker choice Trevin Chow 2026-08-18 20:26:46 -07:00
  • 90f4afb2e1 fix(ce-compound): put the subagent-contract and rendering rules where they fire Trevin Chow 2026-08-18 20:21:37 -07:00
  • 97e6e861d6 fix(ce-work): 48% smaller SKILL.md, references carry the execution manual Trevin Chow 2026-08-18 20:15:01 -07:00
  • 3827c94b60 fix(ce-compound): 90% smaller SKILL.md, references carry the phases Trevin Chow 2026-08-18 20:08:50 -07:00
  • 6da9341faf fix(ce-debug): load the test-selection rule before the Phase 2 recommendation tmchow/8kb-ce-debug-pins Trevin Chow 2026-08-18 19:14:34 -07:00
  • 77c9cd8826 fix(ce-code-review): state the <root> substitution as one stage-level condition tmchow/8kb-ce-code-review Trevin Chow 2026-08-18 19:11:12 -07:00
  • efb29c2fc4 test(ce-brainstorm): repoint the relocated pins by load-time Trevin Chow 2026-08-18 18:59:55 -07:00
  • b201ee3f33 fix(ce-brainstorm): move the phase text into six references Trevin Chow 2026-08-18 18:59:48 -07:00
  • 58586d7926 fix(ce-code-review): carry the resolved artifact root into persona prompts; gate the run artifacts Trevin Chow 2026-08-18 18:53:28 -07:00
  • 3680baa381 fix(ce-debug): audit body pins by provenance, relocate what only wording held in the window Trevin Chow 2026-08-18 17:49:56 -07:00
  • 0217f45dca fix(ce-code-review): place each relocated block where it executes, not where it sat Trevin Chow 2026-08-18 17:44:03 -07:00
  • 7bab56fb02 Address PR review feedback (#1470) tmchow/8kb-ce-plan Trevin Chow 2026-08-18 17:43:33 -07:00
  • a7f6eee2c0 fix(ce-code-review): 86% smaller SKILL.md, four step references carry the rest Trevin Chow 2026-08-18 17:34:55 -07:00
  • e0d814b786 Address PR review feedback (#1470) Trevin Chow 2026-08-18 17:32:35 -07:00
  • 59eb21c8dc test(ce-plan): repoint relocated pins by load-time, not wholesale Trevin Chow 2026-08-18 17:21:40 -07:00
  • 8ca01222a0 fix(ce-plan): relocate the workflow phases behind point-of-use required reads Trevin Chow 2026-08-18 17:21:33 -07:00
  • 7120170f8f fix(ce-explain): an accepted improvement handoff fires, it is not just offered tmchow/8kb-ce-explain-2 Trevin Chow 2026-08-18 17:18:44 -07:00
  • 6cbdcee59a fix(ce-explain): state Phase 6's publish stop as a stop, not a runnable sequence Trevin Chow 2026-08-18 17:13:04 -07:00
  • e8b4b783a3 fix(ce-explain): keep the stop classes that must hold without the Phase 6 read Trevin Chow 2026-08-18 17:03:56 -07:00
  • d3ea223239 docs(skill-design): record the ce-debug re-measurement of the same condition Trevin Chow 2026-08-18 17:00:53 -07:00
  • b16a66f1c0 fix(ce-explain): move the Phase 6 destination close to a required-read reference Trevin Chow 2026-08-18 16:57:26 -07:00
  • 090e7b2a82 fix(ce-pov): 66% smaller SKILL.md, references carry the rest (#1440) Trevin Chow 2026-08-18 16:38:53 -07:00
  • 30f1c69c84 Merge origin/main into tmchow/8kb-ce-pov tmchow/8kb-ce-pov Trevin Chow 2026-08-18 16:30:51 -07:00
  • 11e7a550ef fix(ce-pov): body defers failure vocabulary and tier scope to the references that own them Trevin Chow 2026-08-18 16:29:46 -07:00
  • 8bbcc197bd fix(ce-resolve-pr-feedback): judge top-level PR comments, not only review threads (#1467) Trevin Chow 2026-08-18 16:27:12 -07:00
  • 884d57e589 Merge origin/main into tmchow/8kb-ce-pov (OVER_BUDGET set) Trevin Chow 2026-08-18 15:59:30 -07:00
  • 637f026887 Merge remote-tracking branch 'origin/main' into tmchow/resolve-top-level-comments tmchow/resolve-top-level-comments Trevin Chow 2026-08-18 15:58:54 -07:00
  • 1444dfccf2 fix(ce-resolve-pr-feedback): 31% smaller SKILL.md, references carry the rest (#1435) Trevin Chow 2026-08-18 15:58:30 -07:00
  • e0abc03e9f Merge remote-tracking branch 'origin/tmchow/8kb-ce-resolve-pr-feedback' into tmchow/resolve-top-level-comments Trevin Chow 2026-08-18 14:52:37 -07:00
  • db16478660 fix(ce-test-browser): resolve the dev-server port with a bundled script instead of a prose-emitted line (#1468) Trevin Chow 2026-08-18 14:50:54 -07:00
  • 42dcabfcd5 fix(ce-test-browser): read the .env port as the first digit run, not every digit tmchow/test-browser-port-line Trevin Chow 2026-08-18 14:41:27 -07:00
  • ea696430a4 fix(ce-resolve-pr-feedback): state what makes a non-thread item actionable Trevin Chow 2026-08-18 14:25:08 -07:00
  • b3661f6e3f Merge origin/main into tmchow/test-browser-port-line Trevin Chow 2026-08-18 14:24:35 -07:00
  • f0bb804a77 fix(ce-test-browser): resolve the dev-server port with a bundled script instead of a prose-emitted line Trevin Chow 2026-08-18 14:22:20 -07:00
  • cf098d2304 fix(ce-debug): body enum names fixed-not-pushed like the producer and consumer (#1466) Trevin Chow 2026-08-18 14:21:40 -07:00
  • 1135d54adb fix(ce-babysit-pr): surface the PR author's own non-thread feedback Trevin Chow 2026-08-18 14:09:35 -07:00
  • a6f062d544 fix(ce-debug): drop the stale enum cardinality from the pipeline status line tmchow/ce-debug-body-enum Trevin Chow 2026-08-18 14:08:03 -07:00
  • 2ef62a8120 Merge commit '1ee6244666de2cb8be858d6bcc3e86c54acb6016' into tmchow/resolve-top-level-comments Trevin Chow 2026-08-18 13:56:47 -07:00
  • f411d5c418 fix(ce-resolve-pr-feedback): judge top-level PR comments, not only review threads Trevin Chow 2026-08-18 13:54:50 -07:00
  • b3b8d94ae2 fix(ce-debug): body enum names fixed-not-pushed like the producer and consumer Trevin Chow 2026-08-18 13:53:10 -07:00