cflmy
6705e2e6b7
fix(agent): require explicit local prompt for model workflow steps
2026-11-18 09:56:49 +08:00
cflmy
e8f1a3e126
fix(agent): deduplicate equivalent numbered and labeled user replies
2026-11-18 09:56:16 +08:00
cflmy
6f929dc2da
feat(agent): publish safe partial fan-out recovery through Agent ABI
2026-11-18 09:55:19 +08:00
cflmy
dd22180269
feat(agent): recover failed read-only children only, preserving successful subtask results
2026-11-18 09:55:11 +08:00
cflmy
770dee5199
fix(agent): prevent EffectLedger atomic temp collisions under parallel operations
2026-11-18 09:54:06 +08:00
cflmy
f6612813e3
fix(agent): remove resume test global cwd races and make atomic temp names unique
2026-11-18 09:53:58 +08:00
cflmy
0cd64a7450
fix(agent): persist partial child outcomes and require verified process success before workflow commit
2026-11-18 09:53:02 +08:00
cflmy
1e1f05c087
feat(agent): register ABI workflow state transitions and cancellation
2026-11-18 09:51:59 +08:00
cflmy
06e61efc51
feat(agent): implement plugin-only durable step workflow with deterministic next/commit/cancel and tests
2026-11-18 09:51:51 +08:00
cflmy
629ad8ed39
feat(agent): tolerate short deterministic small-model selection phrases
2026-11-18 09:43:59 +08:00
cflmy
02a4f2d20d
fix(agent): skip YAML setext lookalikes in document-scope parser
2026-11-18 09:40:50 +08:00
cflmy
633436e40f
fix(agent): compare owned JSON values with dereferenced optional identity values
2026-11-18 09:39:53 +08:00
cflmy
c7c0f66d1f
feat(agent): register typed concurrent task result aggregator
2026-11-18 09:38:33 +08:00
cflmy
d5ab9daebb
feat(agent): classify parallel child successes, failures and unknown outcomes
2026-11-18 09:38:26 +08:00
cflmy
e8d6dc92fc
feat(agent): register user-input pause/reply gate primitives
2026-11-18 09:37:44 +08:00
cflmy
afe31cb2a8
feat(agent): durable user-input gate with idempotent reply, revision and session checks
2026-11-18 09:37:35 +08:00
cflmy
f63b7ef48d
feat(agent): register scoped context projection in native plugin
2026-11-18 09:37:02 +08:00
cflmy
393558ca97
feat(agent): register scoped context projection in native plugin
2026-11-18 09:36:59 +08:00
cflmy
0d9bd99fc0
feat(agent): add Markdown heading-scoped bounded file/class/function prompt projection with tests
2026-11-18 09:36:52 +08:00
cflmy
f56deffefe
test(agent): isolate EffectLedger tests without process-global cwd races
2026-11-18 09:33:28 +08:00
cflmy
7fdd97f675
feat(agent): classify invalid model output and bound local recovery
2026-11-18 09:30:33 +08:00
cflmy
b3867c46db
feat(agent): register native choice parser in plugin ABI
2026-11-18 09:30:10 +08:00
cflmy
1a39a53d93
feat(agent): add fail-closed small-model choice parser and tests
2026-11-18 09:30:03 +08:00
cflmy
c0dc18433b
Fix 2.7.0 engineering release compatibility and preserve accepted Ubuntu orig
2026-11-10 17:53:08 +08:00
cflmy
d9aef957f7
Release 2.7.0: executable handbooks and native user-site workflows
2026-11-10 17:03:06 +08:00
cflmy
5080136d65
Make Web an executable handbook with recursive Markdown layouts
2026-11-08 21:50:01 +08:00
cflmy and Cursor
4018e9928a
fix(ai): stream usage events and accept free-form plan replies
...
Request stream_options.include_usage, emit usage SSE events for collect_usage,
propagate plan tokens through run, and treat unknown prose decisions as DONE.
Co-authored-by: Cursor <cursoragent@cursor.com >
2026-10-28 11:56:51 +08:00
cflmy and Cursor
11bb7ef277
feat(agent): wire EffectLedger through mutating run_tool
...
Prepare/start/commit write and execute tools with idempotent dedupe, and refuse
blind retry while ledger state is unknown or interrupted mid-flight (MA-T03).
Co-authored-by: Cursor <cursoragent@cursor.com >
2026-10-27 13:44:45 +08:00
cflmy and Cursor
6bd90bb2d0
feat(agent): ship output channels, ContextPackV2, and effect ledger
...
Open AgentOutputV1/ContextPackV2 capability bits and add EffectLedgerRecordV1
so CheckpointV2 write auto-resume only unlocks on a safe watermark.
Co-authored-by: Cursor <cursoragent@cursor.com >
2026-10-27 13:28:03 +08:00
cflmy and Cursor
4683f2e067
feat(agent): ToolSpec registry with authorize-before-call
...
Add ToolSpecV1 register/search/describe/prepare so CALL and native paths share one schema gate; unauthorized and invalid-arg prepares fail with zero executions.
Co-authored-by: Cursor <cursoragent@cursor.com >
2026-10-27 13:18:55 +08:00
cflmy and Cursor
5912164039
feat(agent): CheckpointV2 atomic resume and AgentErrorV1
...
Replace bare resume writes with temp/fsync/rename, mark legacy loads as manual_review, and expose bounded read-only retry decisions while keeping write auto-resume closed without a ledger.
Co-authored-by: Cursor <cursoragent@cursor.com >
2026-10-27 11:03:39 +08:00
cflmy and Cursor
b8265a92aa
fix(web): stop short/read/control from filling execute slots
...
Isolating every short invoke on the 2-slot worker pool caused no_slot under
normal traffic; keep short paths on parent (execute already off-process) and
background-reap after observer disconnect so zombies release slots.
Co-authored-by: Cursor <cursoragent@cursor.com >
2026-10-27 08:56:56 +08:00
cflmy and Cursor
9af929971c
fix(web): disconnect≠cancel and isolate short/read/control lanes
...
Observer SSE close no longer cancels execute by default; short/read/control
bypass parent INVOKE_LOCK via isolated admission, with read_lane=1 and gold samples.
Co-authored-by: Cursor <cursoragent@cursor.com >
2026-10-27 08:44:07 +08:00
cflmy and Cursor
3cabe53da2
release: v2.1.4 isolated HTTP execute (ADR 0009)
...
Default execute/sse-live/stream to Supervisor worker pool; MR-01 gold;
/_runtime/*; bump CLI/wasm/ext to 2.1.4.
Co-authored-by: Cursor <cursoragent@cursor.com >
2026-10-26 18:33:33 +08:00
cflmy and Cursor
1b11fe063a
fix(web): fail-fast when invoke lock is busy; document single-interp limit
...
Long sse-live chats held INVOKE_LOCK indefinitely and starved short routes
(UI freeze). Timed try-lock returns interpreter busy → HTTP 503; ADR 0008
records that HTTP is multi-goroutine while Marqdo evaluation stays single-flight.
Also lands plan/delta streaming and soft map guards used by qdagent chat.
Co-authored-by: Cursor <cursoragent@cursor.com >
2026-10-26 13:27:53 +08:00
cflmy and Cursor
b792f44556
feat(web): auto sse-live when chat body has stream:true
...
Upgrade json invoke routes to the M02 EventBus forwarder for OpenAI-compatible
stream requests, and normalize LLM delta frames to message.delta.payload.delta
so tokens flush before run.completed.
Co-authored-by: Cursor <cursoragent@cursor.com >
2026-10-26 09:47:05 +08:00
cflmy and Cursor
c7c329341b
feat(workbench): live selection smoke, F02 agent e2e, DocumentView CLI
...
Hoist llm.ask prompt binding to avoid elevated `p`; prefer reasoning_content
when content is empty; point live/tests at in-tree MARQDO_EXT; add USTC live
gates and mock agent.step selection+resume evidence; view document --out for
read-only browser embed checks. Design stays Partial (not Verified).
Co-authored-by: Cursor <cursoragent@cursor.com >
2026-10-25 23:05:28 +08:00
cflmy and Cursor
8cf5288481
feat(workbench): F14 coalesce, snapshot manifest, and sse drain evidence
...
Hold analysis locks across compile for concurrent same-snapshot reads,
emit SnapshotManifestV1 for app retention, and add Go drainForward tests.
Co-authored-by: Cursor <cursoragent@cursor.com >
2026-10-25 11:13:53 +08:00
cflmy and Cursor
af0bda2e40
test(workbench): add F02 subtask/agent skeleton and cancel bridge evidence
...
Cover selection env propagation, agent skeleton without llm.llm, step-kinds
DocumentView read purity, and Go cancel-token bridge smoke.
Co-authored-by: Cursor <cursoragent@cursor.com >
2026-10-25 11:02:25 +08:00
cflmy and Cursor
090b026547
test(workbench): harden F-matrix evidence and demote Verified claim
...
Add dual-endpoint LLM isolation, usage revision replace, cancel/F08/F11
asserts, and mark the design status Partial until remaining F gaps close.
Co-authored-by: Cursor <cursoragent@cursor.com >
2026-10-25 11:01:02 +08:00
cflmy and Cursor
03396afd02
feat(workbench): land M00–M06 runtime contracts and read-only view
...
Add capabilities/selection/usage/DocumentView/locator/RunContext with
honest probes, LLM messages+cancel, sse-live cancel bridge, and CI
workbench_contracts gate; record design evidence for verified slices.
Co-authored-by: Cursor <cursoragent@cursor.com >
2026-10-25 10:55:08 +08:00
cflmy and Cursor
dc853fb580
release: bump to 2.1.2 (EK production chain + View V0–V3 bar)
...
Ship evidence/closure invalidation, verify/adjudicate, reuse_execute and
agent bind/publish gates, document.extract, KnowledgePresentation + SSE
page_token, and upload max_bytes with the 45-test offline gate.
Co-authored-by: Cursor <cursoragent@cursor.com >
2026-10-24 22:52:03 +08:00
cflmy and Cursor
0e1bfb3359
feat: complete BoundCall dual-backend and reuse bind gate (G02)
...
Share resolve_for_exec across Tree/Bytecode and semantic projections,
emit bound_calls/dynamic_calls, and require eng_reuse bind_targets before
execute so discovery scores cannot be treated as handles.
Co-authored-by: Cursor <cursoragent@cursor.com >
2026-10-23 23:06:34 +08:00
cflmy and Cursor
ecbdd18303
feat: executable-knowledge S0–S1 gate plus effect and MLSP reuse subset
...
Establish author baseline fixtures and offline tests, make SemanticProject
the sole authority for graph edges, classify named steps, and tighten
local reuse so unknown effects are not pure and MLSP binds by SymbolId.
Co-authored-by: Cursor <cursoragent@cursor.com >
2026-10-23 22:52:06 +08:00
cflmy
fb135ae2f8
refactor(plugins): 抽出 marqdo-plugin-sdk — ABI 前言从 5 份变 1 份
...
新增工作区成员 crates/marqdo-plugin-sdk:ABI 类型、HOST_* 静态回调、
host_strdup/set_out/set_err/parse_args/reply/host_query_json(_with)/
take_host_string/init_plugin,以及生成四个 #[no_mangle] 导出的
marqdo_plugin! 宏。ABI 常量、符号名、结构体布局一律未变。
5 个插件改为依赖 SDK + 一行宏调用:
agent 1237 → 929 (-308)
quantum 1045 → 835 (-210)
linalg 688 → 524 (-164)
llm 405 → 278 (-127)
demo 170 → 73 (-97)
SDK 新增 232 行 → 净减约 674 行
迁移中发现并修正三次"编译通过但语义已变":
1. 注册项抽取漏 11 项 —— 正则 [a-z_0-9]+ 匹配不到 mcp_server::agent_mcp_serve、
sandbox::* 这类路径限定函数名;改为允许 :: 后重抽。
已用脚本逐项比对 (名称, 参数串, 函数) 三元组:agent 63/63、quantum 41/41、
linalg 31/31、llm 2/2、demo 2/2,全部 0 差异。
2. mcp_client 被误判死模块(删 327 行后才发现 agent_mcp_client 正注册它)→ 恢复。
3. 量子插件两个"未使用"函数删除时连带吃掉 density_from_amps/flat_to_nested
→ 恢复原文件并改为 #[allow(dead_code)]。
验证:cargo build --release(含 5 个插件)零警告;
cargo test --release 609 passed / 0 failed;
量子金样实跑 p00-ok / p11-ok / ping-ok。
文档:doc/design/change-plugin-sdk.md
2026-10-22 15:49:13 +08:00
cflmy
13012cac0e
refactor: 精简 — 删除归档的 Rust web 插件(14,235 行)
...
审计结论:plugins/web-rust-archive/ 不是 Cargo workspace 成员、默认不构建、
无任何脚本/代码引用,官方 web 插件早已是 Go 版(plugins/web/ + libweb.so)。
仅作为"参考"留在树里,占 Rust 代码量约 18%。
删除内容:
- plugins/web-rust-archive/(31 个文件,14,235 行 Rust,528K)
- 更新三处指向:Cargo.toml 注释、plugins/web/README.md、ext-web-go-rewrite.md,
说明历史保留在 git 中
- CHANGELOG Unreleased 记录此次精简
验证:cargo build --release 通过;cargo test --release 565 passed / 0 failed。
这是「不要让冗余代码毁坏 marqdo 设计」的第一步;后续按审计报告继续:
HostFn 元数据表合一(~400L)、plugin ABI 共用 SDK(~550L)、
catalog 复用 knowledge 助手(~140L)、view 路由表化(~130L)、
as_text 19→1(~90L)、view/html 字符串拆分与去重(~300L)。
2026-10-22 15:09:31 +08:00
cflmy
3330a0fc66
chore: 还原 go.mod 被构建副产物改动的依赖分类
...
上一个提交误带入 plugins/web/go.mod 变更:构建 libweb 时用了 -mod=mod,
Go 把 jackc/pgx/v5 与 yuin/goldmark 从 indirect 提升为 direct。
这与本次文档提交无关,属于构建副产物,还原为提交前状态(内容与版本不变,
仅 require 块归类复原)。
2026-10-22 11:59:50 +08:00
cflmy
42a1a46383
docs: 链接式调用 [函数名](参数) 评估 + 开发环境脚本 dev-env.sh
...
两件事:
1. doc/design/link-style-call-evaluation.md(Proposed · 待决策)
评估用户提议的 Markdown 链接式调用 [函数名](参数…)。
结论:语法优雅,但与既有特性「链接索引 [key](map)」形状完全相同,无法可靠消歧。
实测证据:
- [a](m) → 链接索引,取成员,返回 1(既有特性,v0.3 明确推广)
- [greet](alice) → 被当链接索引处理,报 undefined variable alice
两者区别只在「greet 是函数还是键名」,需要符号表信息,而语法判定发生在
解析期(跨模块时未知)。若靠「已知函数表」消歧,会把清晰的
missing map key 变成误导性的 unknown function。
给出四个受限方案并对比,推荐 L3:用「] 与 ( 之间是否有空格」消歧
- [key](coll) 紧贴 → 链接索引,语义一字不动(零回归)
- [callee] (args) 有空格 → 调用,参数显式成组,顺带解决 D8 空格函数名
并附 8 条验证计划。未改任何代码。
2. scripts/dev-env.sh(新增,可 source 复用)
本机测试环境缺失导致基线本来就有失败,本次装好并固化:
- 约束1:$HOME 不可写 → zig/rust 缓存显式重定向
- 约束2:/tmp 是 noexec tmpfs → TMPDIR 必须指向可写且可执行目录
(这是 Go 链接报 "signal: aborted"、zig 报 AccessDenied 的真正原因)
- 提供 rust 1.81 / go 1.25 / python 3.12 / gcc shim(zig, glibc 2.36)
环境修复前后:
- 无 python3:231 passed / 2 failed
- 加 python3:233 / 0(lib)
- 加 go + libweb:gold 284 / 0
- 全量:554 passed / 0 failed
libweb.so 构建要点:Go cgo 需要 gcc,用 zig cc 充当
(/workspace/qdagent/.tools/zigcc/bin/gcc,过滤 Go 传入的
-Qunused-arguments,该参数会让 zig 驱动 abort)。
变更文档同步追加 §8 记录环境安装过程与前后对比。
2026-10-22 11:59:14 +08:00
cflmy and Cursor
8353ba38fe
Add Streamable HTTP MCP and harden web /v1 invoke auth.
...
agent_mcp_serve gains http transport; invoke injects Bearer/X-Api-Key and maps
unauthorized to 401; openai/agent field access prefers json.get over footnotes.
Co-authored-by: Cursor <cursoragent@cursor.com >
2026-10-16 17:33:24 +08:00
cflmy and Cursor
1908bf1a53
fix(ai): soft LLM transport errors and strip ext/ai json glue
...
Host/plugin HTTP failures return structured finish|status=error instead of
aborting the VM; agent/llm/openai/智能体 result maps use GFM tables and
table.put. SKILL gains positive copy-paste examples so agents stop mirroring
legacy json.set chains.
Co-authored-by: Cursor <cursoragent@cursor.com >
2026-10-16 14:27:45 +08:00