46 Commits

Author SHA1 Message Date
lilleman cb3135bc8c Answer principles round 1: 1 🔴, 7 🟡 (0 misread); every effect is handed in as an argument, no test patches a global, and the traces principle scopes tools that run on their user's machine
CI / tests (pull_request) Successful in 3m43s
2026-11-19 12:42:03 +01:00
lilleman eb64154384 Answer prose round 1: 0 🔴, 7 🟡 (0 misread); name the transcripts and review files the tool takes, scope the readers' claim, and rewrite the flagged comments
CI / tests (pull_request) Successful in 2m59s
2026-11-17 02:33:15 +01:00
lilleman b9876c17fe Gzip loop-claude's archive and the review folders 30 days after each file was last written, and read gzipped files by their plain name
CI / tests (pull_request) Failing after 2m59s
2026-11-17 02:00:32 +01:00
lilleman 2f0708f390 Answer prose round 1: 0 🔴, 5 🟡 (0 misread); the decision entry and the help text reworded with the measurement redone, the title names the judge runs, and the hash comments say prompt and reply 2026-11-17 01:15:35 +01:00
lilleman 2600fbb254 Answer stability round 1: 1 🔴, 5 🟡 (0 misread); imports sorted, the replay cuts the task and the rules from its grounds and fails on a missing heading, outcome parameters keyword-only, the mtime order dropped 2026-11-17 01:13:39 +01:00
lilleman b5afb4c36a Kill the mutants of the judge run archive that its tests missed 2026-11-17 00:51:39 +01:00
lilleman e1d2eae7f4 Archive each judge run's prompt and json reply beside the verdict it gave, and replay from the archived prompt
CI / tests (pull_request) Failing after 2m45s
2026-11-16 23:42:36 +01:00
lilleman 64eb8fc7d7 Name the text-file helpers for what they hold, and keep what counts as a quote beside the judge 2026-11-13 02:49:07 +01:00
lilleman 14e6756a53 Answer prose round 1: 0 🔴, 6 🟡 (0 misread); rewrite each <verb> nothing to the negation rule, tighten the new docstrings and comments 2026-11-12 20:33:32 +01:00
lilleman 7e29dc0014 Answer principles round 1: 0 🔴, 4 🟡 (0 misread); drop typing.cast and subTest loops from new code, state the price list's two uses, record the subTest decision 2026-11-12 20:18:54 +01:00
lilleman 5c37027046 Answer architecture round 1: 1 🔴, 8 🟡 (0 misread); one reader per message, prices and Spent block moved, one table of larv-review commands 2026-11-12 20:05:52 +01:00
lilleman fa54ed46b1 Answer prose round 1: 0 🔴, 7 🟡 (0 misread); rename Record to Annotation, and reword the decision, the help and the changelog
CI / tests (pull_request) Successful in 2m37s
2026-11-12 10:25:30 +01:00
lilleman 72d9a9f652 Answer product-owner round 1: 0 🔴, 2 🟡 (0 misread); print the report before exiting 1 on an unreadable archive, naming the file 2026-11-12 10:25:30 +01:00
lilleman 1fc7bf0f52 Kill the surviving mutants of the spend, the setup version and the reversal count, and keep a subagent's type on one line 2026-11-12 10:03:47 +01:00
lilleman 3c4cedd389 Record with each archived verdict the setup that ran it, what each model spent by kind, and the human's reversals
CI / tests (pull_request) Failing after 2m12s
2026-11-12 09:57:56 +01:00
lilleman 12d70586ed Answer principles round 1: 0 🔴, 1 🟡 (0 misread); test the trials' imports from bin, pass one loop value through loop/, type the Exempt reasons
CI / tests (pull_request) Successful in 2m20s
2026-11-10 22:20:48 +01:00
lilleman 8d9a39de1c Group bin's root modules into packages named by purpose, and make rank_todo.py a launcher of release_plan/ranking.py 2026-11-08 20:35:52 +01:00
lilleman 2f6ff46f99 Count the judge's cost in the routing figures 2026-11-04 20:32:23 +01:00
lilleman c583900199 Answer product-owner round 1: 0 🔴, 1 🟡 (0 misread); stamp the transcript and trust lines through say 2026-10-28 18:47:39 +01:00
lilleman 387a5ce506 Answer architecture round 1: 0 🔴, 2 🟡; finish "record", record the three meanings as a decision, name rule-hook's check once 2026-10-23 10:54:24 +02:00
lilleman cb48719493 Answer architecture round 1: 0 🔴, 4 🟡; move verdicts out of cycle, derive RECORDED, type the archived worker as Record, drop stale help names 2026-10-22 22:00:09 +02:00
lilleman 9e4aebb3ff Answer prose round 1: 0 🔴, 4 🟡; fix this_chunks's and segments's cases, move the decision's validity to its first paragraph, name ladder.py and reflow the help, trim the README sentence 2026-10-22 01:03:03 +02:00
lilleman 26f1ef018d Describe the transcript archive's layout where it lives, and make help, decision and comment prose easier to find 2026-10-22 00:58:28 +02:00
lilleman 5b536afdf9 Answer stability round 1: 0 🔴, 5 🟡; tell the resumed worker its stepped rung, rename the step for rulings, build NAMES_A_WORKER from the verdict constants, and test the capped answer and a failed interrupt archive 2026-10-17 22:37:11 +02:00
lilleman fc24e9bae2 Step the worker up as the judge quotes an answer when the run also cites an escalation trigger 2026-10-17 22:24:37 +02:00
lilleman 602a5fac16 Answer principles round 1: 1 🔴, 2 🟡; name the worker's rung on an interrupt and archive its step-up at once, step from the taken-in item, drop every escalation after the first step of a cycle, and state stepped's rules once 2026-10-17 22:10:33 +02:00
lilleman 1c503e8818 Step the worker up as the judge interrupts when the run also cites an escalation trigger 2026-10-17 21:57:11 +02:00
lilleman d9d007fa7a Answer prose round 1: 0 🔴, 1 🟡; list no kinds of rejected reply in the changelog, say rejected throughout, and rewrap the help paragraph 2026-10-16 23:27:24 +02:00
lilleman 23dac2c1d1 Archive each judge reply the loop rejects, with why and the reply 2026-10-16 23:11:29 +02:00
lilleman 798bf505b2 Answer prose round 1: 0 🔴, 8 🟡; say whom the worker asks, give the rung-request rule its own sentence, cite where the triggers are listed, reorder the changelog 2026-10-15 22:26:43 +02:00
lilleman 501fc8c606 Resume a worker one rung up next cycle when it ends its run asking for a higher rung and the judge finds the escalation trigger it cites 2026-10-15 21:55:11 +02:00
lilleman 015e39ae97 Answer stability round 1: 0 🔴, 6 🟡; name no rung without evidence, test the prompt's task, premises and records and the replay's edges, and name a copy and its outcomes file in one module 2026-10-15 09:16:11 +02:00
lilleman 16d762a5f3 Answer architecture round 2: 0 🔴, 2 🟡; archive each verdict's judge rung, list the archive through transcript_archive alone, and say how the replay's retry differs from the loop's 2026-10-15 08:49:22 +02:00
lilleman 16aba4cd98 Answer architecture round 1: 2 🔴, 13 🟡; test the replay's main and report, share one judge run and the archive format, pin the reference judge, route only replayed models, and say why the judge climbs 2026-10-15 08:39:10 +02:00
lilleman acdba1abf4 Answer architecture round 1: 0 🔴, 12 🟡; break the import loop, share the interrupt hand-over and formats, measure under a lock 20 cycles past the last measurement, go on off a terminal, and add a reset 2026-10-15 00:49:02 +02:00
lilleman b1168ca8c5 Measure routing figures every 20th cycle and raise each call for a routing change once 2026-10-15 00:37:01 +02:00
lilleman f52228d9ad Answer prose round 1: 0 🔴, 6 🟡; correct what the help, changelog, decision and item 136 claim, and tighten the routing comments 2026-10-14 23:32:04 +02:00
lilleman bdfe20a65c Answer stability round 2: 0 🔴, 1 🟡; test a failed step-up archive, rename NAMES_A_WORKER, and take the product-owner's wording for help and the worker prompt 2026-10-14 23:28:09 +02:00
lilleman cb9a949cd5 Answer stability round 1: 0 🔴, 7 🟡; archive each step up as it happens, leave the bundled rule to the judge's schema, and test the routing edges mutants survived 2026-10-14 23:23:42 +02:00
lilleman f26e331314 Answer architecture round 3: 0 🔴, 1 🟡; test routing's plan and archive sources, read the newest outcomes file for a rerun's rung, and note the archive reads Rung's wording 2026-10-14 23:12:04 +02:00
lilleman d6e019444d Answer architecture round 2: 2 🔴, 3 🟡; route from the ladder by rung, keep a stepped-up rung across reruns and reloads, test the judge's bundled range, and file item 136 for routing on the item taken 2026-10-14 23:06:54 +02:00
lilleman e2e7a0688f Route each chunk's worker by its item's R + C on one ladder of model and effort rungs 2026-10-14 22:48:31 +02:00
lilleman 7e13778191 Answer prose round 1: fold round-count's changes into the larv-review changelog entry, revise the decisions citing item 83 and stating the floor, reword the help, docstrings and seat-prompts placeholder, drop the archive layout comment that restates --help, and test the past-the-cap line 2026-10-09 15:20:11 +02:00
lilleman b2a277b642 Answer architecture round 4: count a repeated round number, move larv-review's record commands to record.py, give the UTC stamp and the blockquote one home each, check the history's sample entry, and state why a split and a fall call more seats 2026-10-09 15:01:29 +02:00
lilleman dde06fc465 Answer architecture round 1: give loop-claude's prompt text one home, split larv-review's plan trigger and ruling base out, share todo.md's table and the project key, and name the scoring run, the checkout roots and the test env for what they are 2026-10-09 14:43:19 +02:00
lilleman b1913ed65d Restructure bin/ so the project reads at the comprehension floor 2026-10-09 14:12:16 +02:00