Repository navigation
feat(autoresearch): update to 0.2.0 - #188
luochang212 wants to merge 6 commits into
Conversation
|
✅ Internal validation passed. A ZCode maintainer will review the change next; after acceptance, release checks run separately and this PR is closed automatically. |
|
✅ Internal validation passed. A ZCode maintainer will review the change next; after acceptance, release checks run separately and this PR is closed automatically. |
…nd viewing protocol (still 0.2.0)
|
✅ Internal validation passed. A ZCode maintainer will review the change next; after acceptance, release checks run separately and this PR is closed automatically. |
…commit (still 0.2.0) - detectPlateau: deterministic benchmarks switch plateau to no-progress counting (window with no run beating the pre-window global best) instead of the 1% magnitude threshold that misfired on fine-grained metrics; non-deterministic semantics unchanged - init_experiment: optional founding_commit folds pre-existing uncommitted work into a segment: prefixed founding commit, re-establishing "after init, tree == HEAD" so keeps no longer ride along and discards restore the founding state; default path keeps the advisory warning - README (en/zh) updated; plugin suite 256 pass / 0 fail (7 platform skips)
|
✅ Internal validation passed. A ZCode maintainer will review the change next; after acceptance, release checks run separately and this PR is closed automatically. |
|
✅ Internal validation passed. A ZCode maintainer will review the change next; after acceptance, release checks run separately and this PR is closed automatically. |
Full-tree sync of the source plugin/ tree (still 0.2.0), covering the 2026-11-15 regression batch: frozen-write guard dual endpoint/quotes, memory-inject bounded clipping, dashboard Host validation and segment escaping, exclude-file newline healing, rollback cleanup failures surfaced, measurement semantics and plateau threshold tightening, checks negative exit codes, hook settlement, clear broadcast, JSON-RPC protocol hygiene, config.json top-level defense, launcher /usr/bin candidate — each with its field-incident regression tests; plus the matching plugin README (EN/CN) updates for launcher candidates, Console style and clear conditions. Plugin suite: 266 pass / 0 fail (7 platform-specific skips); upstream validate.py clean. No marketplace metadata change.
|
✅ Internal validation passed. A ZCode maintainer will review the change next; after acceptance, release checks run separately and this PR is closed automatically. |
Plugin type (required) / 插件类型（必填）
Affected plugin name(s) / 涉及的插件名称：
autoresearch(Standard plugin)What / 改动内容
This pull request updates the community plugin
autoresearch(registered via #1) from 0.1.0 to 0.2.0 inexternal_plugins/autoresearch/and bumps the matchingmarketplace.jsonentry. The directory is a full sync of the source repo'splugin/tree at v0.2.0, including deletions (the retiredhooks/permission-gate.ts) and additions (scripts/launcher.mjs,mcp/lib/{bash,frozen,harness,spill}.ts,hooks/examples/before/self-check.sh). A follow-up sync commit adds the loop refinements landed right after the v0.2.0 tag moved (deterministic plateau counting;init_experimentfounding commit); version stays 0.2.0. No marketplace metadata changes beyond the version; icon and category are unchanged.本 PR 将社区插件
autoresearch（由 #1 注册）从 0.1.0 更新至 0.2.0，同步external_plugins/autoresearch/全树并升级marketplace.json条目版本号。同步包含删除（退役的hooks/permission-gate.ts）与新增（scripts/launcher.mjs、mcp/lib/{bash,frozen,harness,spill}.ts、hooks/examples/before/self-check.sh）；v0.2.0 tag 移动之后落地的两项循环精化（deterministic 平台期计数；init_experiment段首提交）由后续同步提交并入，版本仍为 0.2.0。除版本号外市场元数据不变，图标与 category 不变。Highlights of 0.1.0 → 0.2.0 (grouped; the source repo tags each fix with the field incident or regression test that motivated it):
scripts/launcher.mjswith the registered command back to barenode(PATH → nvm → standard install locations, Node ≥ 24 gate, signal exit codes passed through, stdout kept clean for JSON-RPC). This raises the runtime requirement from Node ≥ 22 to Node ≥ 24.run_experimentpersists an observation snapshot andlog_experimentmust reconcile against it (one successful log per run; a non-crash metric must equal the observed value);checks_failedrequires a recent failed checks verdict; secondary-metric constraints now read observed values, not caller transcription; delta in the response anchors to the retained line (the same gate the keep audit enforces).asi.infra) and excluded from doom-loop detection; deterministic benchmarks suppress the noise-ratio confidence; keep criterion is aligned to the current-best ratchet across protocol text, audit message, and response..zcode) excluded from keep commits and rollbacks; run lifecycle always settles (bounded fallback finalize); git failures fail closed; spill buffer hygiene; JSON-RPC error responses for parse errors and unknown methods; many field-incident regressions covered by tests.init_experimentgains an opt-infounding_committhat folds pre-existing uncommitted work into asegment:-prefixed founding commit, re-establishing "after init, tree == HEAD" so keeps cannot ride along and discards restore the founding state (default path keeps the advisory warning verbatim).0.2.0 主要变更（分组；每个 fix 在源仓库均有对应的现场事故或回归测试记录）：
scripts/launcher.mjs启动，注册命令回到裸node（解析顺序 PATH → nvm → 标准安装位，主版本门 Node ≥ 24，signal 退出码透传，stdout 保持 JSON-RPC 纯净）。运行要求从 Node ≥ 22 提升到 Node ≥ 24。run_experiment持久化观测快照，log_experiment强制对账（一次 run 一次成功落账；非 crash 的 metric 必须等于观测值）；checks_failed需要最近一次失败判定背书；次级度量约束改读观测值而非调用方转写；响应中的 delta 锚定守擂线（与 keep 审计同一道门）。harness_edit/frozen_attempt）；基础设施故障分离（asi.infra）并从 doom-loop 检测中排除；确定性基准抑制噪声比置信度；keep 判据（对当前最优）在协议文本、审计报错与响应三处口径对齐。.zcode）不再被 keep 提交或回滚波及；run 生命周期必然结算（有界 fallback finalize）；git 故障 fail-closed；溢出缓冲卫生；JSON-RPC 解析失败与未知方法返回标准错误码；大量现场事故均已固化为回归测试。init_experiment新增可选founding_commit，把开段前未提交工作折入segment:前缀段首提交，重建「init 后树 = HEAD」不变量（keep 不再夹带、discard 可恢复段首内容），缺省路径 ADR-11 警告逐字不变。、编辑存档、汇总边界修复。
Testing / 测试方式
node --test tests/*.test.ts— 256 pass, 0 fail (7 platform-specific skips), including regressions for each field incident listed above. Source repo CI runs lint + typecheck + tests on ubuntu / macOS / Windows (Node 24).python3 scripts/validate.py— the only 2 problems reported are pre-existing on currentmainand unrelated (plugins/excalidrawandplugins/openpencilnot registered inmarketplace.json); nothing aboutautoresearch. A targetedbuild_zipofexternal_plugins/autoresearch(61 files) succeeds;git diff --checkclean.Version and registration / 版本与市场注册
autoresearch0.2.0; version bumped identically inexternal_plugins/autoresearch/.zcode-plugin/plugin.jsonand the rootmarketplace.jsonentry; 0.1.0 never reused."skills": "./skills"(same convention applied to thegithubplugin earlier)."skills": "./skills"声明（与此前github插件同一约定）。Dependencies, network, permissions, side effects / 依赖、网络、权限与副作用
.auto/measure.sh) and correctness checks (.auto/checks.sh); automatic git operations (keep commits, non-keep reverts,.auto/exempt); installs 4 hooks (Stop continuation, PreToolUse frozen-write guard, UserPromptSubmit memory injection, SessionStart — the PermissionRequest gate from 0.1.0 is removed);export_dashboardserves a local 127.0.0.1 HTTP page; writes.auto/session state in the project directory. No network access beyond localhost; no model/API keys.Licensing / 许可与来源
assets/icon-sources.json; mechanism research credits (karpathy/autoresearch, pi-autoresearch) noted in the READMEs. No new third-party material in this update.