@tea-agent/loop-agent 0.44.0-next.7 → 0.44.0-next.9

package/docs/templates/frontend-implementation-contract.schema.json

@@ -386,26 +386,18 @@
         "additionalProperties": false,
         "required": [
           "id",
-          "type",
           "commandLabel",
           "file",
           "requirementIds",
-          "uiStates"
+          "uiStates",
+          "mode",
+          "commandId"
         ],
         "properties": {
           "id": {
             "type": "string",
             "minLength": 1
           },
-          "type": {
-            "enum": [
-              "static",
-              "unit",
-              "component",
-              "integration",
-              "mock"
-            ]
-          },
           "commandLabel": {
             "type": "string",
             "minLength": 1
@@ -436,6 +428,34 @@
               "type": "string",
               "minLength": 1
             }
+          },
+          "mode": {
+            "enum": [
+              "static",
+              "behavior"
+            ]
+          },
+          "commandId": {
+            "type": "string",
+            "minLength": 1
+          },
+          "scope": {
+            "enum": [
+              "unit",
+              "component",
+              "integration"
+            ]
+          },
+          "evidenceLevel": {
+            "enum": [
+              "unit",
+              "mounted",
+              "real-integration"
+            ]
+          },
+          "boundary": {
+            "type": "string",
+            "minLength": 1
           }
         }
       }

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@tea-agent/loop-agent",
-  "version": "0.44.0-next.7",
+  "version": "0.44.0-next.9",
   "type": "module",
   "bin": {
     "loop-agent": "bin/loop-agent.js",

package/skills/frontend-design-review/SKILL.md

@@ -15,17 +15,17 @@ Do not rediscover sources, search specification roots or invent connector result
 
 ## Authoritative Terminal
 
-Submit findings one at a time with `record_design_finding` using stable IDs.
-Use `replace:true` for revisions. The terminal automatically includes saved findings;
-do not repeat their array. Correct rejected calls, then commit one successful terminal:
+Use `record_design_finding` with stable IDs; correct rejected calls before committing one terminal, which includes saved findings:
 
 - `approve_design` when no blocking design defect remains;
 - `request_design_changes` when a blocker exists. It must include a typed
   `issueCategory`, at least one `evidenceRef`, and non-empty findings (saved or supplied).
 
-The typed terminal is the only authoritative outcome. Do not require or emit a
-`VERDICT:` line, JSON verdict, revision patch, or a second final-design-review
-verdict. Requested changes end this run; recovery owns the bounded follow-up.
+Correct findings before completing their scope. Identical replay is safe; `replace:true` requires an existing ID and an unfinished scope.
+
+Optional `finding.repairAssertions` may name frozen obligations, e.g. `[{"check":"state-present","token":"loading","source":"AC-1"}]`. Invalid assertions retain the blocker; passing checks still requires semantic review.
+
+Only the typed terminal controls the outcome. No `VERDICT:` line, JSON verdict or revision patch. Requested changes end this run; recovery owns follow-up.
 
 Check every execution group against its members' complete outcomes. Permissions,
 thresholds and error paths remain independent ACs; a shared VT must prove them all.
@@ -47,12 +47,9 @@ Knowledge-base absence is advisory if relevant rules from either openspec
 specification directory were read and applied. Block skipped fallback, unresolved
 conflict, or unresolved UI decisions.
 
-## Method And Human-readable Output
-
-Map criteria to steps/files/states/checks, audit paths/evidence, classify Blocking
-or Advisory, and never edit files.
+## Method
 
-### Spec Evidence Rules
+Audit criteria, steps, files, states and checks; classify Blocking/Advisory. Never edit files.
 
 Read only the canonical contract (or its indexed, hash-bound sections) and any exact
 evidence paths bound in its upstream context. Successful paired reads are the only proof that a specification was read;

package/skills/frontend-design-review/references/review-checklist.md

@@ -9,9 +9,8 @@
 ## Project Fit
 
 - Reuse components, hooks, API helpers, mocks, schemas, router patterns, tokens, and theme rules.
-- Cite knowledge base, `openspec/schemas/`, `openspec/project-specs/`, and `ai_workspace/` as parallel sources; failed or empty
-  knowledge queries must still search `<repoRoot>/openspec/schemas/`, `<repoRoot>/openspec/project-specs/`, and `<repoRoot>/ai_workspace/`.
-- Record source status, query terms, paths/headings, conflicts, authorized deps, and allowed paths for both.
+- Audit frozen source status and exact contract/readSet references. Empty knowledge queries require upstream fallback evidence or a recorded gap. Do not search specification roots; report missing required evidence to its upstream owner.
+- Check the recorded source status, query terms, exact paths/headings, conflicts, authorized deps, and allowed paths.
 - Component selection conformance: every UI purpose must be declared in the contract `uiComponentChoices[]`. `specified` must carry a non-empty `specReference { path, section, line }` hitting a supported openspec/ai_workspace spec; `new` must declare a deviation rationale with `specReference` null; a spec-defined component must not be silently replaced or self-invented without a declared deviation.
 
 ## Interaction / Quality
@@ -37,7 +36,7 @@
 
 ## Verdict Matrix
 
-- Request revision for coverage gaps, unsafe scope, unauthorized deps, unresolved required interaction, missing required verification, skipped local specification fallback (`openspec/schemas/`, `openspec/project-specs/`, or `ai_workspace/`), unsafe/missing Mock strategy, or Mock evidence presented as real integration.
+- Request revision for coverage gaps, unsafe scope, unauthorized deps, unresolved required interaction, missing required verification, missing required upstream specification fallback evidence, unsafe/missing Mock strategy, or Mock evidence presented as real integration.
 - Request revision for component selection non-conformance: spec-defined components silently replaced or self-invented without a declared deviation, `uiComponentChoices` missing for UI-visible work, or a `uiComponentChoices` `specReference.path` not cited in the `openspec-citations` block / not actually read.
 - Knowledge-base unavailable but relevant OpenSpec or `ai_workspace/` rules applied is advisory only.
 - Optional cleanup that cannot affect acceptance is advisory.

package/skills/frontend-review/SKILL.md

@@ -10,7 +10,7 @@ references:
 # Frontend Review
 
 Use `contracts/frontend-review-context.json`, its hash-bound contract index and diff,
-and only files named by the diff. Read the findings guide and capacity diagnostic;
+and the supplied hash-bound test-source/report scopes (including unchanged tests). Read the findings guide and capacity diagnostic;
 use indexed sections when advised. Do not reopen task sources, plans, OpenSpec,
 design-review prose or `ai_workspace/`.
 
@@ -25,6 +25,10 @@ rejected calls, then commit one successful terminal:
   write, or unmet acceptance criterion. It must include a typed `issueCategory`,
   at least one `evidenceRef`, and non-empty findings (saved or supplied).
 
+Save and correct every finding before completing its scope. An identical finding is an idempotent replay; `replace:true` requires an existing ID and cannot modify a completed scope/terminal.
+
+Optional `finding.repairAssertions` may name frozen obligations, e.g. `[{"check":"state-present","token":"loading","source":"AC-1"}]`. Invalid assertions retain the blocker; passing checks still requires semantic review.
+
 Only the typed terminal controls the outcome; no `VERDICT:` line or JSON verdict.
 Requested changes end this run; the recovery controller owns follow-up.
 

package/skills/loop-agent/references/command-reference.md

@@ -119,6 +119,14 @@ export CODE_AGENT_PI_BACKEND=cli-only    # 紧急回滚：纯 CLI 路径
 
 SDK 回归或 SDK 可选依赖不可用时用 `cli-only` 诊断。CLI fallback 路径须与现有 workflow 行为兼容。
 
+### 提升完成 run 的证据
+
+```bash
+loop-agent promote-run <task-id> --run-id <run-id>
+```
+
+`promote-run` 把已完成 DAG 或 one-shot run 的确定性证据提升进任务产物（`修改记录.md`、`验证结果.md`），不重新执行业务步骤。仅在目标 run 已终态后使用；`dag reconcile-tasks` 的 `promotion-drift` 提示会给出该命令与对应 `run-id`。它只写任务产物，不改源码、不替代验收结论。
+
 ### 初始化与旧项目更新
 
 用户中文表达到命令的映射（与目标项目 `AGENTS.md` 的“自然语言入口路由”保持一致）：

package/dist/worker/console/static/assets/channel-5_os7t_x.js

@@ -1 +0,0 @@
-import{ag as o,ah as n}from"./mermaid.core-DF6mMPJ9.js";const t=(a,r)=>o.lang.round(n.parse(a)[r]);export{t as c};

package/dist/worker/console/static/assets/classDiagram-JCYQIIEL-loz0n18W.js

@@ -1 +0,0 @@
-import{s as a,c as s,a as e,C as t}from"./chunk-GF5L2VYU-AY2ZSAx2.js";import{_ as i}from"./mermaid.core-DF6mMPJ9.js";import"./chunk-5VM5RSS4-C5BFY2Po.js";import"./chunk-XXDRQBXY-Bp1xW3T4.js";import"./chunk-KBJHAD2P-gpvvriOp.js";import"./chunk-2GRJ4B5K-BCrcxEh2.js";import"./index-DWKejNVJ.js";var n={parser:e,get db(){return new t},renderer:s,styles:a,init:i(r=>{r.class||(r.class={}),r.class.arrowMarkerAbsolute=r.arrowMarkerAbsolute},"init")};export{n as diagram};

package/dist/worker/console/static/assets/classDiagram-v2-OCEON4UE-loz0n18W.js

@@ -1 +0,0 @@
-import{s as a,c as s,a as e,C as t}from"./chunk-GF5L2VYU-AY2ZSAx2.js";import{_ as i}from"./mermaid.core-DF6mMPJ9.js";import"./chunk-5VM5RSS4-C5BFY2Po.js";import"./chunk-XXDRQBXY-Bp1xW3T4.js";import"./chunk-KBJHAD2P-gpvvriOp.js";import"./chunk-2GRJ4B5K-BCrcxEh2.js";import"./index-DWKejNVJ.js";var n={parser:e,get db(){return new t},renderer:s,styles:a,init:i(r=>{r.class||(r.class={}),r.class.arrowMarkerAbsolute=r.arrowMarkerAbsolute},"init")};export{n as diagram};