“harness”的搜索结果

指导将Everything Claude Code的技能和钩子接入DeepSeek Harness，涵盖配置生成、运行验证、更新，以及技能或钩子未被识别时的排查。

• 向DSH添加ECC技能
• 将钩子接入配置档
• 排查技能或钩子未被识别的问题

Evaluate and improve the policy instructions that guide an AI agent using scored coding tasks. Run a learning cycle to compare candidates, or use the default dry run to inspect tasks and estimated costs without model calls.

中文简介正在准备中。当前显示英语简介。

• Improving agent policies against a scored task set
• Estimating learning costs before running models
• Comparing candidates on held-out evaluation tasks

Assess a repository's readiness for an AI agent execution framework across five dimensions. Produce JSON or a table with readiness scores, estimated cost per run, and scaffold status, without changing repository files.

中文简介正在准备中。当前显示英语简介。

• Assessing readiness before adopting an agent execution framework
• Comparing readiness scores across five dimensions
• Checking estimated run costs and scaffold readiness

Turn a Bifrost PR or GitHub issue into API regression tests that catch recurring bugs. Review existing coverage, add cases to the Postman collection, and validate collection processing and provider filtering.

中文简介正在准备中。当前显示英语简介。

• Add regression coverage for a merged PR
• Turn an issue's reproduction steps into API tests
• Test streaming and non-streaming responses

Create a fixed JSON evaluation suite from a repository's existing tests and check its structure for errors. Use the suite with harness-evolve to compare variants against consistent tasks, independently of the repository's normal test command.

中文简介正在准备中。当前显示英语简介。

• Create a fixed evaluation suite from existing tests
• Check evaluation suite structure in CI
• Compare variants across commits using consistent tasks

Create a starter structure for a custom AI agent harness, the setup that supports agent execution. Choose a template and host, preview the plan, then explicitly confirm generation in a new directory outside the current project.

中文简介正在准备中。当前显示英语简介。

• Creating a coding agent harness from a template
• Preparing starter setups for different agent hosts
• Previewing generated files' destination before writing

Delegate scoped repository analysis or implementation to DeepSeek Harness without opening its interactive interface. Collect its final text, check the exit status, and inspect repository changes when the task requires read-only work.

中文简介正在准备中。当前显示英语简介。

• Diagnose a failing test without editing files
• Delegate a scoped repository investigation
• Request implementation in a specified project

Validate OpenHarness through real model API calls on an unfamiliar project. Tests cover conversations across multiple turns, actual tool execution, and combined features, with guidance for distinguishing setup problems from product failures.

中文简介正在准备中。当前显示英语简介。

• Validate agent behavior with real model API calls
• Check context reuse across multiple conversation turns
• Test hooks, skills, and plugins together

Guides deployment of Open Harness on Vercel, from collecting credentials and setting environment variables to checking the first session. Separates a basic setup from a fuller setup with GitHub repository access.

中文简介正在准备中。当前显示英语简介。

• Deploying a personal copy of Open Harness
• Identifying credentials needed for a basic deployment
• Adding GitHub repository access

Run a Meta-Harness-style optimization loop NATIVELY — automatically search over the scaffolding around a FIXED base model (memory, retrieval, context construction, prompt templates, summarization, tool-selection logic) by proposing candidate variants, scoring each on a cheap deterministic eval, and keeping a Pareto frontier of quality vs cost — using native Agent / Workflow / loop tools instead of a standalone Python harness. Use this whenever the user wants to optimize, evolve, tune, distill, or search over a harness, scaffold, prompt system, memory or retrieval policy, context-assembly code, or summarizer while keeping the model fixed; whenever they mention Meta-Harness, harness optimization, scaffold evolution, automatic prompt/memory optimization, an evolutionary or Pareto search over candidate implementations, or "make the harness/agent better without retraining"; and whenever the gain must come from the code AROUND the model rather than the model weights. Reproduces the Meta-Harness paper's method natively, with no claude_wrapper.py and no metered solver API.

中文简介正在准备中。当前显示原文说明。

xp-harness の skill / agent を新規作成・改修するとき、または xp-harness そのものの設計判断 (取り込む / 取り込まない / 翻訳する) を考えるときに必ず発火させる skill。xp-harness の中核思想 (最終形の理想像 = エージェント群と利用者で XP の開発チームを成す、価値で導きつつ規律装置最小注入、エンジニアとして振る舞う、ペアプロ哲学、中央集権より decentralized、outside-in 例外なし、対話と自走の境界、INVEST の取捨、共創を目指す対話) を context に inject し、新規 skill / agent の設計判断や既存 skill / agent の改修判断を、xp-harness 思想と整合させるためのもの。

中文简介正在准备中。当前显示原文说明。

Plan and review new Omnigent harness integrations, the adapters connecting agent runtimes to the platform. Compare SDK and subprocess approaches with native interface wrappers, using capability matrices and checklists for implementation and testing.

中文简介正在准备中。当前显示英语简介。

• Plan a new agent runtime integration
• Compare SDK and native integration requirements
• Review approval, streaming, and session support

Audit AGENTS.md, rules, skills, and their references for broken paths or commands. Optional independent agent reviews assess redundant instructions and practical usefulness, producing reports that identify what to retain, review, or consider trimming.

中文简介正在准备中。当前显示英语简介。

• Checking paths and commands cited in agent instructions
• Finding redundant rules across skills and instructions
• Assessing which instruction files change agent behavior

Run a Meta-Harness-style optimization loop NATIVELY — automatically search over the scaffolding around a FIXED base model (memory, retrieval, context construction, prompt templates, summarization, tool-selection logic) by proposing candidate variants, scoring each on a cheap deterministic eval, and keeping a Pareto frontier of quality vs cost — using native Agent / Workflow / loop tools instead of a standalone Python harness. Use this whenever the user wants to optimize, evolve, tune, distill, or search over a harness, scaffold, prompt system, memory or retrieval policy, context-assembly code, or summarizer while keeping the model fixed; whenever they mention Meta-Harness, harness optimization, scaffold evolution, automatic prompt/memory optimization, an evolutionary or Pareto search over candidate implementations, or "make the harness/agent better without retraining"; and whenever the gain must come from the code AROUND the model rather than the model weights. Reproduces the Meta-Harness paper's method natively, with no claude_wrapper.py and no metered solver API.

中文简介正在准备中。当前显示原文说明。

Scan declared MCP servers and tools for permission, policy, and dependency issues without executing them. Produce JSON or a Markdown table, and set a severity threshold that determines whether a CI check fails.

中文简介正在准备中。当前显示英语简介。

• Review declared MCP server and tool permissions
• Check policy and dependency issues in MCP configuration
• Fail CI when findings reach a chosen severity

Clarify responsibilities in a TeamHarness team, including planning, task execution, result acceptance, and runtime administration. Use the role boundaries to identify who owns each action and how participants should handle sensitive information.

中文简介正在准备中。当前显示英语简介。

• Assigning responsibilities across team roles
• Checking ownership of task review and acceptance
• Defining work boundaries for remote members

Organize TeamHarness project and task files in shared workspaces, synchronize changes, and submit deliverables. It also guides publishing files to Matrix chat rooms and explaining file access through other requester channels.

中文简介正在准备中。当前显示英语简介。

• Organizing project and task deliverables
• Synchronizing shared workspace files
• Publishing an extra file to a Matrix room

Understand SenseNova Team Harness, a workspace you host yourself for collaboration between people and local AI agents. Learn its core concepts, startup requirements, and workflow for assigning tasks, sharing input files, and reviewing versioned deliverables.

中文简介正在准备中。当前显示英语简介。

• Assess whether SenseNova Team Harness fits a team
• Learn how shared spaces, tasks, and deliverables relate
• Get started with a self-hosted workspace

The harnesses a campaign keeps, read outside a run: which there are and where each came from, what each is made of and what moved from its parent, what each was measured to be, trying one on a task, the audit before a promotion and the blind review at a handover, whether a part earns its place, and changing what makes a harness or what a seat is told about one. Use when the operator asks which harness is current, what a Builder changed, whether a harness is ready to promote or hand over, to run a kept harness on a task, or to change the harness contract. Reading one run is whole-run-investigation's, wording a claim attribution-and-proof's.

中文简介正在准备中。当前显示原文说明。

检查通过进化调优的提示词配置，验证其结构并查看生成的系统提示词，也可对运行记录中的失败类型进行分类，帮助在引入代理前审查配置。

• 采用前检查已调优的提示词配置
• 在CI中验证配置变更
• 对运行记录中的失败进行分类

使用10个已知漏洞和9个容易混淆的正常样例，比较四种安全检测方式，并将检出率、误报、修复测试、复现、安全风险输出和成本等结果汇总为JSON。

• 比较安全检测方式
• 跟踪检出率与误报变化
• 发现定期评估中的不合格结果

以只读方式检查MCP连接与操作相关的威胁，按威胁整理发现并报告最高严重程度，可输出表格或JSON，帮助准备发布前的安全审查材料。

• 开展发布前的安全检查
• 准备提交给安全团队的报告
• 在定期审计中检查威胁变化

以只读方式检查代码仓库，汇总仓库类型、推荐的智能体角色配置、风险、测试可信度和发布准备情况，并可根据风险阈值进行持续集成判定。

• 构建自定义智能体运行框架前检查仓库
• 比较定期报告以发现配置变化
• 按风险阈值进行持续集成判定

Experiment with the policies that govern an AI agent, such as planning, memory, and tool use. Test one policy change per variant in a sandbox, an isolated environment, and promote only candidates with measured improvements.

中文简介正在准备中。当前显示英语简介。

• Finding policy changes that improve low evaluation scores
• Exploring initial agent settings for a new domain
• Comparing manual tuning with an evolved baseline