Commit Graph

  • 29dca5a005 chore: release main release-please--branches--main github-actions[bot] 2026-08-24 19:38:33 +00:00
  • b20c29d7aa fix(cross-model): drop empty Codex usage files and unblock heartbeat teardown (#1533) main Trevin Chow 2026-08-24 12:37:15 -07:00
  • 20b2ece3a2 fix(cross-model): give the heartbeat its own process group tmchow/debug-issue-1531 Trevin Chow 2026-08-24 11:19:37 -07:00
  • e2f71bb03d fix(cross-model): unblock heartbeat teardown on SIGTERM Trevin Chow 2026-08-24 11:04:49 -07:00
  • c8bb04948c fix(ce-code-review): omit empty adversarial-codex-usage.json Trevin Chow 2026-08-24 10:44:21 -07:00
  • 871996c4e3 chore: release main (#1528) compound-engineering-v3.23.3 github-actions[bot] 2026-08-24 00:23:41 -07:00
  • 5ac4c01a42 fix(ce-code-review): collect async reviewer returns (#1530) Trevin Chow 2026-08-24 00:19:33 -07:00
  • 5d5b78431a fix(ce-babysit-pr): defer to GitHub for trunk drift on managed stacks (#1529) Trevin Chow 2026-08-24 00:17:31 -07:00
  • d4119f0ff4 fix(ce-code-review): settle terminal reviewer outcomes (#1530) tmchow/debug-issue-1523 Trevin Chow 2026-08-24 00:07:54 -07:00
  • 72fea5d5f4 fix(ce-babysit-pr): key managed readiness on stack_blocker, not re-derived freshness tmchow/fix-babysit-pr-base Trevin Chow 2026-08-24 00:01:08 -07:00
  • 4476adf852 fix(ce-code-review): collect async reviewer returns Trevin Chow 2026-08-23 23:53:36 -07:00
  • bd0049ba64 fix(ce-babysit-pr): defer to GitHub for trunk drift on managed stacks Trevin Chow 2026-08-23 23:48:41 -07:00
  • d05e3951cc fix(lfg): invoke ce-code-review from the host catalog path (#1527) Trevin Chow 2026-08-23 23:34:46 -07:00
  • 0572602944 fix(lfg): keep SKILL.md under the Codex 8000-byte bound tmchow/fix-ce-code-review-catalog-path Trevin Chow 2026-08-23 23:28:55 -07:00
  • d0a03a162b fix(lfg): invoke ce-code-review from the host catalog path Trevin Chow 2026-08-23 23:22:00 -07:00
  • 180fcd21e1 fix(skills): match blocking questions from the current tool list (#1526) Trevin Chow 2026-08-23 23:27:35 -07:00
  • d84b095132 Address PR review feedback (#1526) tmchow/debug-issue-1522 Trevin Chow 2026-08-23 23:20:11 -07:00
  • 66ee6c29eb fix(skills): match blocking questions from the current tool list Trevin Chow 2026-08-23 23:09:40 -07:00
  • f293ca43cb docs(readme): make the README a front door, split detail into linked docs (#1525) Trevin Chow 2026-08-23 23:08:27 -07:00
  • a46c74d322 fix(readme): catch a skill left in two overview rows, and claim only what is enforced tmchow/readme-overhaul-docs-split Trevin Chow 2026-08-23 23:06:51 -07:00
  • 8824762028 fix(docs): state the claude prerequisite and stop over-promising host coverage Trevin Chow 2026-08-23 22:56:56 -07:00
  • 39491d5ca7 fix(docs): point contributor setup at the package test script Trevin Chow 2026-08-23 22:45:23 -07:00
  • b420b2343f fix(readme): enforce the README skill overview instead of trusting convention Trevin Chow 2026-08-23 22:41:25 -07:00
  • 62089639bc Merge commit '3187ab1d79c52a6618e654f77a4fa27d4b6c6818' into tmchow/readme-overhaul-docs-split Trevin Chow 2026-08-23 22:35:13 -07:00
  • 7ac4ca5cfd docs(readme): make the README a front door, split detail into linked docs Trevin Chow 2026-08-23 22:28:17 -07:00
  • 0e1d7bb7b2 docs(readme): add a replayable demo of the compounding loop Trevin Chow 2026-08-23 22:28:09 -07:00
  • 3187ab1d79 docs(solutions): refresh stale learnings against current code (#1524) Trevin Chow 2026-08-23 20:58:48 -07:00
  • c9912bd435 docs(solutions): keep release-path and visual-aid thresholds aligned with code tmchow/ce-compound-refresh-stale Trevin Chow 2026-08-23 20:40:28 -07:00
  • 86e26c89ee docs(solutions): refresh stale learnings against current code Trevin Chow 2026-08-23 20:34:45 -07:00
  • a32c9474c6 docs: refresh cross-model peer learnings (#1521) Trevin Chow 2026-08-23 09:55:35 -07:00
  • 3722ca2683 chore: release main (#1515) compound-engineering-v3.23.2 github-actions[bot] 2026-08-23 09:42:26 -07:00
  • 826f2cc9b2 docs: refresh cross-model peer learnings docs/refresh-cross-model-peer-learnings Trevin Chow 2026-08-23 09:36:31 -07:00
  • 2e10508950 fix(cross-model): bound peer retries and raise review headroom (#1519) Trevin Chow 2026-08-23 09:16:20 -07:00
  • 587dbd1a62 fix(cross-model): centralize retry ownership tmchow/search-la-peer-review Trevin Chow 2026-08-23 03:14:21 -07:00
  • 6d03029e1c fix(cross-model): preflight Python classifier Trevin Chow 2026-08-23 03:03:58 -07:00
  • 567beefd8e fix(cross-model): recognize Codex completion events Trevin Chow 2026-08-23 02:54:33 -07:00
  • 2b49ba2177 fix(cross-model): honor final peer outcomes Trevin Chow 2026-08-23 02:42:35 -07:00
  • 5a2ee46497 fix(cross-model): preserve diagnostic stream boundaries Trevin Chow 2026-08-23 02:38:23 -07:00
  • 07f87e6eac fix(cross-model): validate peer overload inputs Trevin Chow 2026-08-23 02:31:21 -07:00
  • 15115e9126 Recognize structured overload errors Trevin Chow 2026-08-23 02:18:51 -07:00
  • 6dcc69851e Reconcile peer retry disclosure Trevin Chow 2026-08-23 02:09:49 -07:00
  • 5e32ff2d8e Stabilize shared retry budget test Trevin Chow 2026-08-23 01:59:49 -07:00
  • 656716d358 Accept Cursor result envelopes Trevin Chow 2026-08-23 01:57:07 -07:00
  • 96f3ba05d8 Gate all recognized peer statuses (#1519) Trevin Chow 2026-08-23 01:34:21 -07:00
  • 286d65ff20 Require successful peer terminal outcomes (#1519) Trevin Chow 2026-08-23 01:24:22 -07:00
  • c2f2b9d6b8 Unify peer outcome classification (#1519) Trevin Chow 2026-08-23 01:12:56 -07:00
  • e850129ae2 Classify complete overload diagnostics (#1519) Trevin Chow 2026-08-23 00:52:40 -07:00
  • 9c431447a7 Reject failed peer terminal envelopes (#1519) Trevin Chow 2026-08-23 00:45:36 -07:00
  • 5934937afd Handle multi-line 529 diagnostics (#1519) Trevin Chow 2026-08-23 00:39:52 -07:00
  • d0eb7cba55 Address second-round PR feedback (#1519) Trevin Chow 2026-08-23 00:32:06 -07:00
  • 56cb13aa15 docs(ce-skill-work): size a behavioral eval to the decision, not the workflow (#1520) Trevin Chow 2026-08-23 00:22:58 -07:00
  • 1eac7c45c6 Address PR review feedback (#1519) Trevin Chow 2026-08-23 00:19:31 -07:00
  • 339915dcc2 docs(ce-skill-work): the decision probe covers routing, not activation; packs run rows serially tmchow/skill-work-proportionate-eval Trevin Chow 2026-08-23 00:14:32 -07:00
  • aa117ed2b4 docs(ce-skill-work): size a behavioral eval to the decision, not the workflow Trevin Chow 2026-08-23 00:09:51 -07:00
  • 2d77dee3da fix(ce-doc-review): activate product-lens only on a product position (#1517) Trevin Chow 2026-08-23 00:09:11 -07:00
  • 32ae3306e3 fix(cross-model): bound peer retries and raise review headroom Trevin Chow 2026-08-23 00:05:01 -07:00
  • c55368fab5 test(skill-eval-cell): a roster probe requires the TEAM trailer tmchow/doc-review-right-size-a Trevin Chow 2026-08-22 23:45:40 -07:00
  • 553ec1682c fix(ce-doc-review): keep only the product-lens change; drop review-artifact persistence Trevin Chow 2026-08-22 23:39:55 -07:00
  • f91d05b7e4 fix(ce-doc-review): manifest records unit and line counts instead of a depth label Trevin Chow 2026-08-22 23:32:34 -07:00
  • 882f1af325 Revert "fix(ce-plan): record depth in unified-plan metadata" Trevin Chow 2026-08-22 23:31:54 -07:00
  • 3d8da53c2e Revert "fix(ce-plan): keep the shared html-rendering reference identical across its three copies" Trevin Chow 2026-08-22 23:31:54 -07:00
  • d4c1b07168 fix(ce-doc-review): persist any local return whose artifact file is absent Trevin Chow 2026-08-22 22:48:29 -07:00
  • cafa7e263f refactor(skill-eval-cell): one hasBaseline predicate for the A/B decision Trevin Chow 2026-08-22 22:41:24 -07:00
  • 7fa3f2df46 refactor(skill-eval-cell): reuse lastTrailer for the TEAM roster scope Trevin Chow 2026-08-22 22:41:12 -07:00
  • 3ce5855fab fix(cross-model): attest Grok Build as a host and bind its native CLI (#1516) Trevin Chow 2026-08-22 22:34:22 -07:00
  • c75a5ddaa7 Address PR review feedback (#1516) tmchow/investigate-grok-doc-review Trevin Chow 2026-08-22 22:12:41 -07:00
  • c24fe36615 fix(cross-model): attest Grok Build as a host and bind the native Grok CLI Trevin Chow 2026-08-22 22:04:54 -07:00
  • 23629434be test(skill-eval-cell): run the A/B pair for any row that carries its own baseline_ref Trevin Chow 2026-08-22 21:53:35 -07:00
  • 36bc42105b test(ce-doc-review): drop the superseded run-directory slot pin Trevin Chow 2026-08-22 21:52:33 -07:00
  • bbcebf788f fix(review): scope roster grade terms to the TEAM line; orchestrator names the persona artifact path Trevin Chow 2026-08-22 21:52:22 -07:00
  • 04fa65e967 fix(ce-plan): keep the shared html-rendering reference identical across its three copies Trevin Chow 2026-08-22 21:37:48 -07:00
  • 7efdcb35a3 docs(plans): doc-review cost measurable plan Trevin Chow 2026-08-22 21:34:22 -07:00
  • 896c20f236 test(skill-eval-cell): product-lens activation probe rows and a must_not_include grade term Trevin Chow 2026-08-22 21:34:04 -07:00
  • 574aabbc7d fix(ce-doc-review): activate product-lens on an unsettled product position, not plausible alternatives Trevin Chow 2026-08-22 21:31:16 -07:00
  • b262f2a01c fix(ce-plan): record depth in unified-plan metadata Trevin Chow 2026-08-22 21:31:04 -07:00
  • 09b55ebdcd fix(ce-doc-review): create the run directory on every review and persist local persona returns Trevin Chow 2026-08-22 21:29:20 -07:00
  • 6f6c5779d3 fix(skills): right-size ceremony for small work in ce-plan, ce-brainstorm, and ce-work (#1514) Trevin Chow 2026-08-22 20:39:30 -07:00
  • 55b9ed1dfe fix(ce-plan): a markdown save of a chat brief stays the plain save tmchow/ce-scope-detection-tuning Trevin Chow 2026-08-22 19:12:41 -07:00
  • 53799d8a60 fix(ce-plan): route a format-naming save of a chat brief to the Durable path Trevin Chow 2026-08-22 19:06:39 -07:00
  • 9d60d45399 fix(ce-plan): pin Durable on a request worded as a plan, not only a plan file Trevin Chow 2026-08-22 18:59:17 -07:00
  • d7cb2b02c7 docs(plan): note the eval-cell dispatch-observer follow-up Trevin Chow 2026-08-22 12:52:50 -07:00
  • e62480d989 fix(ce-work): carry the mechanical no-watch condition into a project-defined shipping process Trevin Chow 2026-08-22 12:52:43 -07:00
  • b14174f00d fix(skills): condition the brainstorm closing summary on artifact presence, state Direct's authority condition once, and grade the risky-small row on the untouched auth file Trevin Chow 2026-08-22 12:46:08 -07:00
  • 5b7db6dad4 fix(ce-plan): kernel Direct lines match the offer-only handoff contract Trevin Chow 2026-08-22 12:40:47 -07:00
  • 838c237ed5 fix(ce-brainstorm): make the Phase 4 handoff reachable on the Lightweight chat path Trevin Chow 2026-08-22 12:39:28 -07:00
  • a5c23f57a6 fix(skills): state the chat-tier outcomes in both kernels, make Direct offer rather than invoke execution, pin Durable on explicit artifact requests, and reserve saved-brief paths Trevin Chow 2026-08-22 12:37:47 -07:00
  • 963894759d docs(plan): record live-session evidence and the ce-debug follow-up Trevin Chow 2026-08-22 12:32:13 -07:00
  • 9f451d5877 fix(ce-plan): keep task visibility and dispatch-rejection handling on the deepen path; align docs and the chat-brief eval row with the shipped gates Trevin Chow 2026-08-22 12:18:02 -07:00
  • 254a25da5b docs(plan): record probe and Direct rerun results in the eval report Trevin Chow 2026-08-22 12:10:26 -07:00
  • bb69119c77 docs(plan): add the right-size ceremony eval report Trevin Chow 2026-08-22 11:57:09 -07:00
  • 22061b794c test(skill-eval-cell): grade Direct's state-and-stop branch in the cell; the invoke branch is live delegation Trevin Chow 2026-08-22 11:56:01 -07:00
  • 1fd3a7b00f test(skill-eval-cell): grade the routing probes on the decision text, not a reference name Trevin Chow 2026-08-22 11:15:32 -07:00
  • a5c519f4f9 test(skill-eval-cell): guard the large paths with bounded routing probes instead of full runs Trevin Chow 2026-08-22 11:07:54 -07:00
  • 2d934b903f test(skill-eval-cell): guard the Durable, Standard-brainstorm, and reviewed-ship paths against the right-size gate Trevin Chow 2026-08-22 10:56:18 -07:00
  • 99bff9bd8b docs(plan): describe the description negative as evidence-gated Trevin Chow 2026-08-22 10:29:34 -07:00
  • 1868860967 test(skill-eval-cell): add git_remote so shipping rows reach the push path; pin Direct's ce-work-unavailable stop Trevin Chow 2026-08-22 10:29:06 -07:00
  • 0cb9de072b fix(skills): restate right-size gates as conditions and drop unverified description negatives Trevin Chow 2026-08-22 10:26:22 -07:00
  • 792d019995 test(skills): reuse loaded brainstorm fixtures in the right-size pins Trevin Chow 2026-08-22 10:15:37 -07:00
  • 83844e37c9 fix(skills): right-size ceremony for small work in ce-plan, ce-brainstorm, and ce-work Trevin Chow 2026-08-22 10:13:30 -07:00
  • 925b4ef71c chore: release main (#1513) compound-engineering-v3.23.1 github-actions[bot] 2026-08-22 08:59:27 -07:00