DeepSeek/deepseek-harness
1
0
Fork 0
mirror of https://github.com/deepseek-ai/deepseek-harness synced 2026-08-15 21:04:50 +00:00
Code Issues Packages Projects Releases Wiki Activity

docs(code-mode): clarify post-policy rendering

Browse Source
This commit is contained in:
Tianyi Cui
2026-07-22 02:28:12 +08:00
parent d68126c3fe
commit ccdc6b0ad3
3 changed files with 4 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
.agents/notes/implemented/bug-fix/2026-07-20-code-mode-result-card-completeness.i18n.yaml
View File

@@ -2,5 +2,5 @@
# side as of the last confirmed-consistent state. Both languages carry equal authority;
# after editing either side, bring the other along and re-record with:
# pnpm run verify-translation-pairing --write
2026-07-20-code-mode-result-card-completeness.md: 5fce1c51dbe029f7c109dfdac55444a51f0dd886
2026-07-20-code-mode-result-card-completeness.zh.md: 624d541cea2e5b4402c880bfe00c586bda7df0ee
2026-07-20-code-mode-result-card-completeness.md: 85942a564b7e4ff6768bedb040ae4371f48bb665
2026-07-20-code-mode-result-card-completeness.zh.md: 3e71667de66e017176266fd7f5d82aeabf507c61

2
.agents/notes/implemented/bug-fix/2026-07-20-code-mode-result-card-completeness.md
View File

@@ -12,7 +12,7 @@ Nested Code calls never owned cards, so producing metadata for the outer call so
## Decision
The canonical tool registry pipeline owns the final model-facing outer content. On success, the `run_code` output renderer renders captured logs followed by the return value or the explicit no-output marker. Runtime failures and policy denials are normalized into error content by `ToolRegistry` without invoking that renderer. Post-execute policy and spill may replace content before persistence.
The canonical tool registry pipeline owns the final model-facing outer content. On success, the `run_code` output renderer renders captured logs followed by the return value or the explicit no-output marker. Runtime failures and pre-execution policy denials are normalized into error content by `ToolRegistry` without invoking that renderer. A post-execute block runs after successful rendering and replaces the result with error content; other post-execute policy and spill decisions may replace content before persistence.
`run_code.presentResult` now forwards the final `result.content` into one generic result card. It deliberately omits the title so the pending card retains the program text. The redundant logs-only `presentationMeta` projection is removed: `tool/result.content` is the durable, replayable, post-policy projection and the card's only result-content source.

2
.agents/notes/implemented/bug-fix/2026-07-20-code-mode-result-card-completeness.zh.md
View File

@@ -12,7 +12,7 @@ Status: implemented
## 决策
规范的工具注册表流水线负责最终面向模型的外层内容。成功时，`run_code` 输出渲染器先渲染已捕获的日志，然后渲染返回值或显式的无输出标记。运行时失败和策略拒绝由 `ToolRegistry` 归一化为错误内容，过程中不会调用该渲染器。Post-execute 策略与输出落盘机制可以在持久化之前替换这些内容。
规范的工具注册表流水线负责最终面向模型的外层内容。成功时，`run_code` 输出渲染器先渲染已捕获的日志，然后渲染返回值或显式的无输出标记。运行时失败和执行前策略拒绝由 `ToolRegistry` 归一化为错误内容，过程中不会调用该渲染器。Post-execute 阻断发生在成功渲染之后，并把结果替换为错误内容；其他 post-execute 策略与输出落盘决策可以在持久化之前替换内容。
`run_code.presentResult` 会把最终的 `result.content` 转交给一张通用结果卡片。它有意省略标题，使待完成卡片保留程序文本。多余的仅含日志的 `presentationMeta` 投影被移除：`tool/result.content` 是持久、可回放且经过 post-policy 处理的投影，也是卡片中结果内容的唯一来源。