Fep Agent Hub

CommunityPopular

Cai-aa

fep-agent-hub

Orchestrate and validate evidence-first CAE workflows through the independent FreeCAD, Elmer FEM, and ParaView MCP servers in FEP Agent Hub. Use for FEP repository setup or diagnostics; CAD-to-mesh-to-solver-to-postprocessing automation; steady or transient heat transfer; 2D harmonic transformer induction; 2D plane-stress beams; 2D steady laminar channel flow; 2D transient eddy-current and Lenz-law studies; regression, acceptance reports, or scientifically honest CAE videos built from these verified profiles.

Overview

Publisher	Cai-aa
Repository	`CAE-Agent-Hub`
Skill name	`fep-agent-hub`
Stars	1K
Forks	130
Bundled files	4
License	MIT
Links	Skill source Repository

Markdown instructions
A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.
Works with any LLM
AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.
4 bundled files
Scripts, templates, and references the model can read while it works. Files are read-only and never executed.
Open source
Published by Cai-aa on GitHub. Read the source before you install it.

Installation

Install the Fep Agent Hub AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

Open the app and go to Plugins → Skills.
Choose "Install from GitHub".
Paste the skill folder URL below and confirm.
https://github.com/Cai-aa/CAE-Agent-Hub/tree/main/Skill/fep-agent-hub
Enable the skill in any chat where you want it available.

Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills

git clone --depth 1 https://github.com/Cai-aa/CAE-Agent-Hub.git /tmp/CAE-Agent-Hub
mkdir -p .claude/skills
cp -r /tmp/CAE-Agent-Hub/Skill/fep-agent-hub .claude/skills/fep-agent-hub

Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Fep Agent Hub in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Fep Agent Hub on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Fep Agent Hub is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

Loaded only when it is needed
The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.
Switch models mid-chat
Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

FEP Agent Hub

Run the three MCP servers as independent stages joined by manifests, semantic IDs, native solver artifacts, and explicit physics gates. Preserve auditability over convenience.

Establish the run contract

Locate the FEP Agent Hub repository and read its AGENTS.md, root README, relevant prompt, and relevant report before changing a workflow.
Resolve native executables through OPEN_CAE_CONFIG, environment variables, or configs/open-cae.local.toml. Treat public paths as placeholders.
Choose a unique project name beneath the configured workspace. Keep FCStd, STEP, mesh, SIF, logs, VTU/PVD, images, tables, and evidence inside that project.
Probe all three MCP environments before modeling. Stop if a required native executable or server is unavailable.
Never expose arbitrary shell, Python, executable, macro, MATC, or raw SIF input. Use only fixed runners, structured parameters, and allowlisted analysis profiles.

Select a verified profile

Read references/case-catalog.md when selecting a profile, copying benchmark parameters, or deciding whether the request exceeds verified capability.

Use a listed profile only for the physics and dimensionality it declares. Add a new profile, deterministic benchmark, tests, and gates before claiming a new class of analysis is supported.

Orchestrate the native chain

Read references/orchestration.md before executing or modifying a case.

Follow this order:

Preflight — inspect tool/resource inventories and probe FreeCAD, Elmer/Gmsh, and ParaView.
FreeCAD — create or open the document; create named features; apply only required booleans/transforms; inspect objects; validate geometry; save FCStd; export STEP plus geometry manifest.
Elmer geometry and mesh — create the case with an allowlisted profile; import STEP and manifest; generate Gmsh physical groups; convert with ElmerGrid; inspect SI bounds, entity counts, and semantic mappings.
Elmer physics — set materials, the structured equation profile, excitations, and boundary conditions; generate and validate SIF; run ElmerSolver; check job state, process exit, logs, expected time steps, actual arrays, finite values, and physics gates.
ParaView — start the MCP-owned headless session; open the actual VTU/PVD; inspect dataset type, arrays, associations, ranges, and time values before creating filters; render/export PNG, CSV, animation frames, and PVSM; inspect the final pipeline; stop only the MCP-owned session.
Report — record inputs, versions, mesh, solver status, field extrema, benchmark errors, sensitivity checks, artifacts, hashes, pass/fail decisions, limitations, and the exact meaning of any animation.

Do not skip directly from file creation to visualization. Each stage must pass before its artifact becomes input to the next stage.

Enforce evidence and claims

Read references/validation-and-claims.md before accepting results, publishing percentages, comparing accuracy, or composing video.

Apply these non-negotiable rules:

Treat SUCCEEDED, BLOCKED, and FAILED as distinct outcomes. A contextually unsupported operation should return BLOCKED, not a fabricated artifact.
Never infer solver success from file presence. Require successful process exit, completion markers, absence of fatal diagnostics, expected artifacts, finite/nontrivial fields, and case-specific physics gates.
Preserve semantic identity across FreeCAD semantic_id -> Gmsh Physical ID -> Elmer body/boundary ID -> VTU array -> ParaView proxy/array.
State contract coverage, call success rate, and numerical error separately. A 100% workflow pass rate is not 100% engineering accuracy.
Keep large native results local. Commit only code, prompts, compact metrics, curated evidence, and reports permitted by .gitignore.
Describe transient, harmonic, quasi-static, and steady media honestly. Postproduction motion never upgrades the underlying physics.

Use repository regression scripts

When a deterministic full-case replay is requested, prefer the repository scripts instead of reconstructing long tool sequences:

scripts/mcp_heat_smoke.py
scripts/mcp_transient_heat_smoke.py
scripts/mcp_transformer_smoke.py
scripts/mcp_beam_smoke.py
scripts/mcp_channel_flow_smoke.py
scripts/mcp_lenz_eddy_smoke.py --variants full
scripts/mcp_full_validation.py
scripts/protocol_smoke.py

Run the repository's current prescribed test matrix before release. Report the observed counts from that run rather than copying historical counts into a new claim.

Block unsupported extensions

Return BLOCKED with the missing profile, model, reference data, or validation work when the request requires unverified nonlinear materials, contact, plasticity, structural dynamics, turbulence, free surfaces, compressibility, three-dimensional end effects, magnetic hysteresis/saturation, motion, bidirectional multiphysics, MPI, arbitrary scripts, or industrial-certification accuracy.

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Fep Agent Hub AI skill do?

Why use Fep Agent Hub on TypingMind?

Because you install it once and use it with any model. Fep Agent Hub is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Fep Agent Hub in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/Cai-aa/CAE-Agent-Hub/tree/main/Skill/fep-agent-hub. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Fep Agent Hub?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Fep Agent Hub?

As many as you like. As long as a model supports skills, you can use Fep Agent Hub with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Fep Agent Hub AI skill free?

Yes. It is published on GitHub by Cai-aa under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Gstack Openclaw Ceo Review

garrytan

CommunityPopular

Use when asked to review a plan, challenge a proposal, run a CEO review, poke holes in an approach, think bigger about scope, or decide whether to expand or reduce the plan.

Caveman Stats

Show recorded output and cache-read token usage and mode attribution for the current Claude Code session, or locate the host's native usage report. Trigger: /caveman-stats.

Ci Metrics

Query PyTorch CI, GitHub Actions, HUD, Grafana, and infrastructure metrics. Use when users ask about CI duration, job failures, queue times, workflow trends, runner health, dashboard data, or PyTorch…

Notcrawl

Notion archive: search, sync freshness, pages/databases, Markdown exports, SQL counts, and Notcrawl repo work.

Overview

Installation

Install in TypingMind

Install in another agent

Use it in TypingMind

Works with any AI model

Skill instructions

FEP Agent Hub

Establish the run contract

Select a verified profile

Orchestrate the native chain

Enforce evidence and claims

Use repository regression scripts

Block unsupported extensions

Bundled files

Frequently asked questions

Related AI Skills

Set up your own AI workspace now