@deepseek-ai/dsh-tool-ralph 0.2.0-rc.1 → 0.2.1-alpha.1

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
package/README.i18n.yaml CHANGED
@@ -30,8 +30,8 @@
30
30
  en: 0999ea3a4c9da48c
31
31
  zh: 83013ee36a8689af
32
32
  /deepseek-ai-dsh-tool-ralph/understand-the-implementation/design-concept:
33
- en: a6aea42b6289ec9e
34
- zh: 305461828b33e67a
33
+ en: 3777bac0215c7366
34
+ zh: 18240f47506c67bd
35
35
  /deepseek-ai-dsh-tool-ralph/understand-the-implementation/fixed-script-and-routing:
36
36
  en: fa2cfda2d9d121e6
37
37
  zh: 75720bd6244a2c96
@@ -45,11 +45,11 @@
45
45
  en: 07e071bd44b1df65
46
46
  zh: 26f1cb2eab8c25b0
47
47
  /deepseek-ai-dsh-tool-ralph/understand-the-implementation/source-map:
48
- en: b0cf0927913968fb
49
- zh: 6170b4c8d2529a41
48
+ en: 1bc11234792b4e3f
49
+ zh: 872d2c5f8df37b72
50
50
  /deepseek-ai-dsh-tool-ralph/further-exploration:
51
- en: 9ab6ec527d57aee1
52
- zh: e0258a158b8027ea
51
+ en: 966417f5c0659ef9
52
+ zh: 6c95a000db46ce0f
53
53
  /deepseek-ai-dsh-tool-ralph/model-experience:
54
54
  en: 215e7ba838619b7b
55
55
  zh: a311da3843709f90
@@ -93,8 +93,8 @@
93
93
  en: 8f28959dc87d1f2f
94
94
  zh: c034c5c0bdaf9bba
95
95
  /deepseek-ai-dsh-tool-ralph/known-limitations-and-deferred-work:
96
- en: 922fd33532125ebb
97
- zh: 27d508697ebc7668
96
+ en: 4e3629e8de05fe18
97
+ zh: 5b33af22079a2ca7
98
98
  /deepseek-ai-dsh-tool-ralph/known-limitations-and-deferred-work/dev-note:
99
99
  en: 8f24c317039e1974
100
100
  zh: c6209969ad21f16c
package/README.md CHANGED
@@ -58,7 +58,7 @@ This section explains the fixed-script design and the validation and lifecycle m
58
58
 
59
59
  ### Design concept
60
60
 
61
- The loop is a deployment-owned fixed script: the model supplies data only and cannot alter the loop, provider route, schema, or handoff validation. The tool is an ordinary plugin over `ctx.workflowEngine` and `ctx.subagents` — no Ralph mode or fresh-agent loop is added to `agent-loop`, and the same-session goal domain stays independent. The [harness-level goal-based execution Agent Note](../../../.agents/notes/implemented/feature/2026-07-16-harness-level-loop.md) owns the policy and deferred work.
61
+ The loop is a deployment-owned fixed script: the model supplies data only and cannot alter the loop, provider route, schema, or handoff validation. The tool is an ordinary plugin over `ctx.workflowEngine` and `ctx.subagents` — no Ralph mode or fresh-agent loop is added to `agent-loop`, and the same-session goal domain stays independent.
62
62
 
63
63
  ### Fixed script and routing
64
64
 
@@ -81,7 +81,6 @@ The pending call is a `generic` card titled `ralph` with the immutable objective
81
81
  | File | Role |
82
82
  |---|---|
83
83
  | [`src/index.ts`](src/index.ts) | Plugin entry: fixed script, provider routing, report validation, tool registration |
84
- | — | No runtime invariant companion is published; this model-facing orchestration adapter owns no independent event stream; workflow and subagent owners validate the runs and child lifecycles it starts. |
85
84
 
86
85
  </details>
87
86
 
@@ -97,7 +96,7 @@ Read these pages when the tool-level contract is not enough. They move from the
97
96
  - [PTC workflow engine](../workflow-ptc/README.md) — the engine that executes the fixed script.
98
97
  - [subagent seam](../../subagent/subagent/README.md) — the fresh-child provider contract.
99
98
  - [Goal group](../../goal/goal/README.md) — same-session goal tools for ordinary long-running objectives.
100
- - [Harness-level goal-based execution Agent Note](../../../.agents/notes/implemented/feature/2026-07-16-harness-level-loop.md) — the policy, provider requirements, and deferred work.
99
+ - [historical Harness-level goal-based execution Agent Note](../../../.agents/notes/archived/feature/2026-07-16-harness-level-loop.md) — the policy, provider requirements, and deferred work.
101
100
 
102
101
  -----
103
102
 
@@ -159,10 +158,11 @@ Each fresh child has an independent request cache. The parent result appends aft
159
158
 
160
159
  These limits define what the tool does not yet support. They are current constraints, not a task backlog.
161
160
 
162
- - **Completion is worker self-declaration** — there is no independent evaluator or verifier deciding whether the objective is complete; evaluator policy and evaluator-driven continuation are deferred.
161
+ - **Completion is worker self-declaration** — there is no independent evaluator or verifier deciding whether the objective is complete; evaluator policy and evaluator-driven continuation are deferred. An evaluator needs separately defined inputs, tool access, deterministic checks, provider choice, isolation, and authority before it can certify completion or drive continuation.
163
162
  - **Foreground only** — there is no job id, background collection, process-resume checkpoint, scheduler, or wall-clock start policy.
164
163
  - **The workspace is the only cross-round long-term memory** — one bounded report is the explicit handoff, and uncommitted conversational reasoning disappears with each child.
165
164
  - **One round is one fresh child** — there is no within-round fan-out, model or provider switching, fork context, or model-call-selected provider.
165
+ - **No enforced recursive-Ralph denial** — structural denial is not implemented; prompt guidance does not enforce a recursion ban.
166
166
  - **Ordinary child failure is terminal for the run** — the fixed script reports the failed round and last successful handoff but does not retry; fatal workflow infrastructure failures can end before that state is returned.
167
167
  - **Only round count bounds aggregate effort** — token, price, and elapsed-time budgets are deferred.
168
168
 
package/README.zh.md CHANGED
@@ -58,7 +58,7 @@ kind: "package-reference"
58
58
 
59
59
  ### 设计理念
60
60
 
61
- 循环是部署方拥有的固定脚本:模型只提供数据,无法改变循环、提供方路由、schema 或交接校验。该工具是基于 `ctx.workflowEngine` 与 `ctx.subagents` 的普通插件——不会向 `agent-loop` 添加 Ralph 模式或全新 agent loop,同会话的 goal 领域也保持独立。[Harness 层目标式执行 Agent Note](../../../.agents/notes/implemented/feature/2026-07-16-harness-level-loop.zh.md)拥有策略与暂缓事项。
61
+ 循环是部署方拥有的固定脚本:模型只提供数据,无法改变循环、提供方路由、schema 或交接校验。该工具是基于 `ctx.workflowEngine` 与 `ctx.subagents` 的普通插件——不会向 `agent-loop` 添加 Ralph 模式或全新 agent loop,同会话的 goal 领域也保持独立。
62
62
 
63
63
  ### 固定脚本与路由
64
64
 
@@ -81,7 +81,6 @@ kind: "package-reference"
81
81
  | 文件 | 职责 |
82
82
  |---|---|
83
83
  | [`src/index.ts`](src/index.ts) | 插件入口:固定脚本、提供方路由、报告校验、工具注册 |
84
- | — | 不发布运行时不变式伴生入口;该面向模型的编排适配器不拥有独立事件流;工作流与 subagent 归属方会校验该适配器启动的运行及其子 agent 生命周期。 |
85
84
 
86
85
  </details>
87
86
 
@@ -97,7 +96,7 @@ kind: "package-reference"
97
96
  - [PTC 工作流引擎](../workflow-ptc/README.zh.md)——执行固定脚本的引擎。
98
97
  - [subagent seam](../../subagent/subagent/README.zh.md)——全新子 agent 的提供方约定。
99
98
  - [goal 组](../../goal/goal/README.zh.md)——面向普通长期目标的同会话 goal 工具。
100
- - [Harness 层目标式执行 Agent Note](../../../.agents/notes/implemented/feature/2026-07-16-harness-level-loop.zh.md)——策略、提供方要求与暂缓事项。
99
+ - [历史Harness 层目标式执行 Agent Note](../../../.agents/notes/archived/feature/2026-07-16-harness-level-loop.md)——策略、提供方要求与暂缓事项。
101
100
 
102
101
  -----
103
102
 
@@ -159,10 +158,11 @@ Use the ralph tool ONLY when the direct human explicitly asks for a Ralph loop o
159
158
 
160
159
  这些限制说明该工具尚未支持什么。它们是当前约束,不是任务积压。
161
160
 
162
- - **完成由 worker 自行声明**——没有独立评估器或验证器判断目标是否完成;评估器策略与评估器驱动的延续均暂缓。
161
+ - **完成由 worker 自行声明**——没有独立评估器或验证器判断目标是否完成;评估器策略与评估器驱动的延续均暂缓。评估器须先单独定义输入、工具访问、确定性检查、提供方选择、隔离与授权规则,才能认证完成或驱动延续。
163
162
  - **仅支持前台**——没有 job id、后台收集、进程恢复检查点、调度器或基于挂钟时间的启动策略。
164
163
  - **工作区是唯一的跨 Round 长期记忆**——一份有界报告作为显式交接,每个子 agent 结束后,未提交的对话推理都会消失。
165
164
  - **一个 Round 对应一个全新子 agent**——Round 内没有扇出、模型或提供方切换、fork 上下文或由模型调用选择的提供方。
165
+ - **没有强制拒绝递归 Ralph**——结构性拒绝尚未实现;提示词引导不能强制禁止递归。
166
166
  - **普通子 agent 失败会终止运行**——固定脚本报告失败的 Round 与上一次成功交接,但不会重试;致命的工作流基础设施失败可能在该状态返回前结束。
167
167
  - **聚合工作量仅受 Round 数量限制**——token、价格与耗时预算均暂缓。
168
168
 
package/package.json CHANGED
@@ -1,7 +1,7 @@
1
1
  {
2
2
  "name": "@deepseek-ai/dsh-tool-ralph",
3
3
  "description": "Model-facing fresh-agent Ralph loop over the workflow and subagent seams",
4
- "version": "0.2.0-rc.1",
4
+ "version": "0.2.1-alpha.1",
5
5
  "publishConfig": {
6
6
  "access": "public"
7
7
  },
@@ -27,32 +27,32 @@
27
27
  ],
28
28
  "license": "MIT",
29
29
  "peerDependencies": {
30
- "@deepseek-ai/dsh-agent": "0.2.0-rc.1",
31
- "@deepseek-ai/dsh-llm": "0.2.0-rc.1",
32
- "@deepseek-ai/dsh-subagent": "0.2.0-rc.1",
33
- "@deepseek-ai/dsh-system-prompt": "0.2.0-rc.1",
34
- "@deepseek-ai/dsh-tools": "0.2.0-rc.1",
35
- "@deepseek-ai/dsh-workflow": "0.2.0-rc.1",
36
- "@deepseek-ai/cordis": "~4.0.4"
30
+ "@deepseek-ai/dsh-agent": "0.2.1-alpha.1",
31
+ "@deepseek-ai/dsh-llm": "0.2.1-alpha.1",
32
+ "@deepseek-ai/dsh-system-prompt": "0.2.1-alpha.1",
33
+ "@deepseek-ai/dsh-subagent": "0.2.1-alpha.1",
34
+ "@deepseek-ai/dsh-tools": "0.2.1-alpha.1",
35
+ "@deepseek-ai/dsh-workflow": "0.2.1-alpha.1",
36
+ "@deepseek-ai/cordis": "~4.0.5-alpha.1"
37
37
  },
38
38
  "dependencies": {
39
- "@deepseek-ai/schemastery": "~3.18.4"
39
+ "@deepseek-ai/schemastery": "~3.18.5-alpha.1"
40
40
  },
41
41
  "devDependencies": {
42
- "@deepseek-ai/cordis-plugin-loader": "~1.0.5",
43
- "@deepseek-ai/dsh-agent": "0.2.0-rc.1",
44
- "@deepseek-ai/dsh-agent-loop": "0.2.0-rc.1",
45
- "@deepseek-ai/dsh-agent-loop-testkit": "0.2.0-rc.1",
46
- "@deepseek-ai/dsh-llm": "0.2.0-rc.1",
47
- "@deepseek-ai/dsh-session": "0.2.0-rc.1",
48
- "@deepseek-ai/dsh-subagent-in-process-driver": "0.2.0-rc.1",
49
- "@deepseek-ai/dsh-system-prompt": "0.2.0-rc.1",
50
- "@deepseek-ai/dsh-subagent-spawn-in-process": "0.2.0-rc.1",
51
- "@deepseek-ai/dsh-tools": "0.2.0-rc.1",
52
- "@deepseek-ai/dsh-workflow": "0.2.0-rc.1",
53
- "@deepseek-ai/dsh-workflow-ptc": "0.2.0-rc.1",
54
- "@deepseek-ai/cordis": "~4.0.4",
55
- "@deepseek-ai/dsh-session-projection": "0.2.0-rc.1",
56
- "@deepseek-ai/dsh-subagent": "0.2.0-rc.1"
42
+ "@deepseek-ai/cordis-plugin-loader": "~1.0.6-alpha.1",
43
+ "@deepseek-ai/dsh-agent": "0.2.1-alpha.1",
44
+ "@deepseek-ai/dsh-agent-loop": "0.2.1-alpha.1",
45
+ "@deepseek-ai/dsh-agent-loop-testkit": "0.2.1-alpha.1",
46
+ "@deepseek-ai/dsh-llm": "0.2.1-alpha.1",
47
+ "@deepseek-ai/dsh-session": "0.2.1-alpha.1",
48
+ "@deepseek-ai/dsh-subagent": "0.2.1-alpha.1",
49
+ "@deepseek-ai/dsh-subagent-spawn-in-process": "0.2.1-alpha.1",
50
+ "@deepseek-ai/dsh-subagent-in-process-driver": "0.2.1-alpha.1",
51
+ "@deepseek-ai/dsh-system-prompt": "0.2.1-alpha.1",
52
+ "@deepseek-ai/dsh-tools": "0.2.1-alpha.1",
53
+ "@deepseek-ai/dsh-workflow": "0.2.1-alpha.1",
54
+ "@deepseek-ai/dsh-workflow-ptc": "0.2.1-alpha.1",
55
+ "@deepseek-ai/cordis": "~4.0.5-alpha.1",
56
+ "@deepseek-ai/dsh-session-projection": "0.2.1-alpha.1"
57
57
  }
58
58
  }