Skills Plugins MCP Prompt Model 博客 我的中心
开发编程 #python #automation #api #ai

comsol-multiphysics

Use when connecting Codex to local or remote COMSOL Multiphysics, probing COMSOL installations, running comsolbatch or comsolcompile, using Java API automation, using Python mph session tools converted from the bundled MCP runtime, searching bundled COMSOL manuals, or repairing COMSOL automation failures.

DeepseekModel 官方收录技能 质量 良好 · 64 v1.0.0

获取

https://deepseekmodel.com/api/download.php?id=wfy-op-codex-for-comsol-lumerical-comsol-multiphysics-skill-md&format=skill
下载 .skill 标准格式,含 system_prompt 与 model_config,导入任意 Agent 框架即可使用
.skill 文件中 system_prompt 字段的实际内容。
name comsol-multiphysics description Use when connecting Codex to local or remote COMSOL Multiphysics, probing COMSOL installations, running comsolbatch or comsolcompile, using Java API automation, using Python mph session tools converted from the bundled MCP runtime, searching bundled COMSOL manuals, or repairing COMSOL automation failures. COMSOL Multiphysics Overview Use this skill to operate COMSOL Multiphysics through two preserved backends: comsolbatch / comsolcompile command-line probes and Java API scripts. Python mph session operations converted from the bundled MCP runtime. Keep the scope to solver connection, model/session management, parameter edits, geometry, physics, mesh, study, result export, documentation lookup, and failure repair. Do not add project-specific device templates or paper-specific modeling assumptions to this skill. Workflow Detect the operating system first: On Windows PowerShell, use scripts/probe_comsol.ps1 and scripts/validate_comsol_skill.ps1 . On Git Bash for Windows, use the .sh scripts; they detect MINGW/MSYS/CYGWIN and delegate to the PowerShell scripts with path conversion. On Linux, use scripts/probe_comsol.sh and scripts/validate_comsol_skill.sh ; they run COMSOL CLI commands directly. Decide which backend is needed: Use the command-line backend for installation probes, Java API builders, batch runs, table exports, and robust local smoke tests. Use the Python mph backend for interactive model/session tools, existing model inspection, parameter management, geometry/physics/mesh/study/result utilities, and documentation helpers. Discover local paths before assuming any profile: Prefer explicit user-provided paths. Then check COMSOL_BIN , COMSOL_ROOT , COMSOL_HOME , COMSOL_MPH_PYTHON , PYTHON , and the current PATH . Then scan common COMSOL installation roots for the newest version folder. If no Python environment can import mph and mcp , ask for or create an appropriate environment instead of using a machine-specific path. Load only the needed reference: Java API, batch, geometry, physics, mesh, result, dataset, or table export syntax: references/comsol-automation.md . Example local profile structure: references/solver-profiles.json . Probe before real work unless the same backend already passed in the current session. Normalize the local solver environment before launching binaries. Run the smallest viable operation first: version check, import check, session start, model load, or compile/run/save probe. Store raw stdout/stderr, generated scripts, result files, and probe JSON near the caller's job artifacts, not inside this skill. If a solver call fails, classify the signature, propose the smallest patch, retry once when safe, then report the exact command and artifacts. Quick Commands Command-line probe: $skill = "C:\path\to\comsol-multiphysics" $out = ".\solver_probe_out\comsol" powershell -ExecutionPolicy Bypass -File "$skill\scripts\probe_comsol.ps1" -DryRun -OutDir "$out" powershell -ExecutionPolicy Bypass -File "$skill\scripts\probe_comsol.ps1" -Deep -OutDir "$out" Git Bash or Linux: skill= "/path/to/comsol-multiphysics" out= "./solver_probe_out/comsol" bash " $skill /scripts/probe_comsol.sh" --dry-run --out-dir " $out " bash " $skill /scripts/probe_comsol.sh" --deep --out-dir " $out " The probe auto-discovers common Windows installs and honors COMSOL_BIN , COMSOL_ROOT , COMSOL_HOME , and binaries already available on PATH . Treat the resolved path as session-local evidence, not as a portable profile. Python runtime adapter: $python = "<python-that-can-import-mph-and-mcp>" $skill = "C:\path\to\comsol-multiphysics" & $python "$skill\scripts\comsol_mph_tool.py" list-tools & $python "$skill\scripts\comsol_mph_tool.py" tool model_load --json "{\"file_path\":\"C:\\path\\to\\model.mph\",\"set_current\":true}" & $python "$skill\scripts\comsol_mph_tool.py" resource "comsol://session/info" Use any Python environment that can import mph and mcp . The adapter reuses the bundled runtime under scripts/comsol_mcp_runtime ; it does not reimplement the MCP tools. For stateful model work, use batch --json [...] or scripts/validate_comsol_skill.ps1 so session state stays in one Python process. Preserved Runtime Capabilities The bundled Python runtime preserves the original MCP categories: Session: start, connect, disconnect, status. Model: load, create, save, version, list, set current, clone, remove, inspect, components. Parameters: get, set, list, sweep setup, descriptions. Geometry: create sequences, add common features, boolean operations, CAD import, build, list features. Physics: available interfaces, add common interfaces, boundary setup, materials, multiphysics couplings, feature listing/removal. Mesh and study: mesh creation/info, study solve, async progress, cancellation, solutions, datasets. Results: expression evaluation, scalar evaluation, inner/outer values, data/image exports, export and plot listing. Knowledge: embedded guides, troubleshooting, best practices, PDF module list/status/search. Resources: session info, model tree, model parameters, model physics. Validation Run the bundled validation script from a caller workspace: $skill = "C:\path\to\comsol-multiphysics" powershell -ExecutionPolicy Bypass -File "$skill\scripts\validate_comsol_skill.ps1" ` -SkillDir "$skill" ` -Python "<python-that-can-import-mph-and-mcp>" ` -ModelPath "C:\path\to\model.mph" Git Bash or Linux validation: skill= "/path/to/comsol-multiphysics" bash " $skill /scripts/validate_comsol_skill.sh" \ --skill-dir " $skill " \ --python "/path/to/python" \ --model-path "/path/to/model.mph" -Python and -ModelPath are optional discovery inputs. Without -ModelPath , validation should still check skill structure, COMSOL probe behavior, runtime imports, and adapter registration. With a supplied model, the validation script must not save over the model; it may load, inspect, list parameters, read resources, and remove the model from memory. Common Mistakes Do not edit files inside the COMSOL installation directory. Do not assume python points to an environment with mph ; use a discovered or user-selected interpreter. Do not treat solver exit code 0 as enough; verify the expected artifact or exported table exists. Do not save back to a user-supplied .mph unless the user explicitly requested it. Do not write project-specific geometry or physics assumptions into this reusable skill.
Agent 识别该技能的关键词,点击任意一个即可复制。

该技能未提供触发词。

下载的 .skill 包内含以下字段。
字段 说明
format格式标识(skill/v1)
skill_id技能唯一 ID
name技能名称
version版本号
description技能描述
category所属分类(数组)
trigger_words触发词列表
tags标签列表
source来源标识
source_url来源链接(本页地址)
exported_at导出时间(每次下载生成)
system_prompt系统提示词正文
model_config模型参数:provider / model / temperature / max_tokens / top_p
examples示例
install_guide各平台导入说明(Coze / Dify / Claude / 自定义框架)
同一份技能可按不同平台格式导出。
.skill 标准格式,含 system_prompt 与 model_config,导入任意 Agent 框架即可使用 下载
.skillpro 增强格式,额外含脚本 / 工具 / 依赖 / 钩子占位 下载
.json 纯 JSON 导出,只含 system_prompt 与模型参数 下载
Coze 带 frontmatter 的 Markdown,Coze 平台导入用 下载
Dify Dify DSL,创建应用后直接导入 下载

每日精选 Skill 推荐,免费送到你邮箱

输入邮箱,每天接收一个精选 AI Agent 技能推荐。完全免费,持续更新。

验证码 --

提交后我们会发送一封确认邮件,点击邮件里的链接才会开始收信。

完全免费,取消任意时间。我们不会发送垃圾邮件。