@tea-agent/loop-agent 0.35.0 → 0.35.1-beta.0

package/skills/browser-tools/package.json

@@ -1,19 +1,19 @@
-{
-	"name": "browser-tools",
-	"version": "1.0.0",
-	"type": "module",
-	"description": "Minimal CDP tools for collaborative site exploration",
-	"author": "Mario Zechner",
-	"license": "MIT",
-	"dependencies": {
-		"@mozilla/readability": "^0.6.0",
-		"cheerio": "^1.1.2",
-		"jsdom": "^27.0.1",
-		"puppeteer": "^24.31.0",
-		"puppeteer-core": "^23.11.1",
-		"puppeteer-extra": "^3.3.6",
-		"puppeteer-extra-plugin-stealth": "^2.11.2",
-		"turndown": "^7.2.2",
-		"turndown-plugin-gfm": "^1.0.2"
-	}
-}
+{
+	"name": "browser-tools",
+	"version": "1.0.0",
+	"type": "module",
+	"description": "Minimal CDP tools for collaborative site exploration",
+	"author": "Mario Zechner",
+	"license": "MIT",
+	"dependencies": {
+		"@mozilla/readability": "^0.6.0",
+		"cheerio": "^1.1.2",
+		"jsdom": "^27.0.1",
+		"puppeteer": "^24.31.0",
+		"puppeteer-core": "^23.11.1",
+		"puppeteer-extra": "^3.3.6",
+		"puppeteer-extra-plugin-stealth": "^2.11.2",
+		"turndown": "^7.2.2",
+		"turndown-plugin-gfm": "^1.0.2"
+	}
+}

package/skills/code-review-core/SKILL.md

@@ -1,20 +1,20 @@
----
-name: code-review-core
-description: 用于只读 code review 节点，产出带文件与行号引用的优先级 findings。
----
-
-# Code Review Core
-
-本 skill 用于 reviewer 节点。
-
-## Review 重点
-
-- 优先关注 correctness、行为回归、security、数据丢失与缺失验证。
-- findings 须 grounded 在具体文件与行号。
-- 区分 blocking findings 与次要 maintainability 备注。
-- 检查测试是否证明变更行为，文档是否与面向用户的变更一致。
-- 若无 findings，明确说明，并指出 residual test gap（如有）。
-
-## Output
-
-按 severity 排序列出 findings。仅当无 blocking findings 时，才写 `VERDICT: pass`。
+---
+name: code-review-core
+description: 用于只读 code review 节点，产出带文件与行号引用的优先级 findings。
+---
+
+# Code Review Core
+
+本 skill 用于 reviewer 节点。
+
+## Review 重点
+
+- 优先关注 correctness、行为回归、security、数据丢失与缺失验证。
+- findings 须 grounded 在具体文件与行号。
+- 区分 blocking findings 与次要 maintainability 备注。
+- 检查测试是否证明变更行为，文档是否与面向用户的变更一致。
+- 若无 findings，明确说明，并指出 residual test gap（如有）。
+
+## Output
+
+按 severity 排序列出 findings。仅当无 blocking findings 时，才写 `VERDICT: pass`。

package/skills/codebase-scout/SKILL.md

@@ -1,19 +1,19 @@
----
-name: codebase-scout
-description: 用于只读 scout 节点，在实现前定位现有代码、测试、文档与集成点。
----
-
-# Codebase Scout
-
-本 skill 用于 scout 节点。
-
-## 规则
-
-- 从 repo 指令、task source 与邻近测试入手。
-- 可用时优先 CodeGraph；否则用 `rg` 与聚焦文件阅读。
-- 在提议新抽象前，识别现有 helper 与 ownership 边界。
-- 只返回事实，不做编辑。
-
-## Output
-
-列出相关文件、现有模式、风险，以及实现所需的最小 write surface。
+---
+name: codebase-scout
+description: 用于只读 scout 节点，在实现前定位现有代码、测试、文档与集成点。
+---
+
+# Codebase Scout
+
+本 skill 用于 scout 节点。
+
+## 规则
+
+- 从 repo 指令、task source 与邻近测试入手。
+- 可用时优先 CodeGraph；否则用 `rg` 与聚焦文件阅读。
+- 在提议新抽象前，识别现有 helper 与 ownership 边界。
+- 只返回事实，不做编辑。
+
+## Output
+
+列出相关文件、现有模式、风险，以及实现所需的最小 write surface。

package/skills/grill-me/SKILL.md

@@ -1,10 +1,10 @@
----
-name: grill-me
-description: Interview the user relentlessly about a plan or design until reaching shared understanding, resolving each branch of the decision tree. Use when user wants to stress-test a plan, get grilled on their design, or mentions "grill me".
----
-
-Interview me relentlessly about every aspect of this plan until we reach a shared understanding. Walk down each branch of the design tree, resolving dependencies between decisions one-by-one. For each question, provide your recommended answer.
-
-Ask the questions one at a time.
-
-If a question can be answered by exploring the codebase, explore the codebase instead.
+---
+name: grill-me
+description: Interview the user relentlessly about a plan or design until reaching shared understanding, resolving each branch of the decision tree. Use when user wants to stress-test a plan, get grilled on their design, or mentions "grill me".
+---
+
+Interview me relentlessly about every aspect of this plan until we reach a shared understanding. Walk down each branch of the design tree, resolving dependencies between decisions one-by-one. For each question, provide your recommended answer.
+
+Ask the questions one at a time.
+
+If a question can be answered by exploring the codebase, explore the codebase instead.

package/skills/loop-agent/SKILL.md

@@ -24,6 +24,7 @@ references:
 ## 初始化升级路由
 
 - 写入型「**loop-agent初始化更新**」「loop-agent 初始化更新」「更新 loop-agent 初始化内容」「升级后对齐」统一运行 `loop-agent init upgrade --repo-root . --json`；处理 controller 返回的单文件 `allowedPaths`，以同一 run `--continue` 推进至终态，不停在中间状态。
+- 升级可能因 gitignore 迁移评估暂停：若 `.agents/` 存在已跟踪内容，需先审查是否有应团队共享的自定义内容；存在 staged 变更或 git 查询失败时需先解决 blocker，再重跑评估。建议的 `git rm -r --cached --ignore-unmatch` 只修改 index、保留工作区文件，且由用户执行而非 controller。
 - 「检查初始化更新」「初始化更新校验」「只检查，不要修改」仅运行 `loop-agent init check-update --repo-root . --markdown`，不得创建 run 或写入。
 - facts 位于 `.harness/init-upgrades/<run-id>/`；禁止业务源码和宽泛 glob。默认管理项目 `.opencode/plugins/`、`.pi/extensions/`、`.pi/settings.json`；Pi trust 后加载，默认不读写 `~/.pi/agent/settings.json`。