Agent Skills
› Cai-aa/CAE-Agent-Hub
› abaqus-job
abaqus-job
GitHub用于创建、提交和监控Abaqus有限元分析作业的技能。支持生成输入文件、并行计算配置及作业状态监控，适用于模型就绪后的仿真执行任务。
Trigger Scenarios
运行分析
提交作业
执行模型
生成输入文件
检查作业状态
Install
npx skills add Cai-aa/CAE-Agent-Hub --skill abaqus-job -g -y
SKILL.md
Frontmatter
{
"name": "abaqus-job",
"description": "Create and manage Abaqus jobs. Use when user asks to run the analysis, submit the job, execute the model, or generate input file.",
"allowed-tools": [
"Read",
"Write",
"Edit",
"Glob",
"Grep",
"Bash(abaqus:*)"
]
}
Abaqus Job Skill
This skill creates, submits, and monitors Abaqus analysis jobs. Use it when the model is ready to run.
When to Use This Skill
Route here when user says:
- "Run the analysis", "Submit the job", "Execute the model"
- "Generate input file", "Create INP file"
- "Run in parallel", "Check job status"
Route elsewhere:
- Reading results after completion →
/abaqus-odb - Setting up the model → use other module skills
Prerequisites
Before job submission:
- Model is complete (geometry, material, mesh, BCs, loads, step)
- Model saved to .cae file
- No validation errors
Workflow: Running an Analysis
Step 1: Save the Model
Always save before creating a job. The .cae file must exist.
Step 2: Create the Job
Specify job name and model name. They can differ.
Step 3: Choose Submission Mode
| User Wants | Action |
|---|---|
| Run analysis and wait | Submit with waitForCompletion |
| Generate INP only (no run) | writeInput |
| Run in background | Submit without waiting |
| Run from command line | abaqus job=Name interactive |
Step 4: Wait and Monitor
For interactive submission, monitor status until COMPLETED or ABORTED.
Step 5: Check Results
If COMPLETED, results are in .odb file. If ABORTED, check .msg file.
Key Decisions
Submit vs Write Input?
| Goal | Method |
|---|---|
| Run analysis now | submit() |
| Only create INP file | writeInput() |
| Run later from CLI | writeInput, then abaqus job=Name |
Parallel Processing
| Scenario | Setting |
|---|---|
| Small model / Learning Edition | numCpus=1 |
| Large model, multi-core | numCpus=N, numDomains=N |
| Single machine | mp_mode=THREADS |
| Cluster | mp_mode=MPI |
What to Ask User
If unclear, ask:
- "Ready to run the analysis?"
- "How many CPUs for parallel?"
- "Just need the input file, or run the analysis?"
Output Files
| Extension | Content |
|---|---|
| .odb | Results database (use /abaqus-odb to read) |
| .dat | Printed output (nodal values, summaries) |
| .msg | Solver messages - check this if job fails |
| .sta | Status file (increment progress) |
| .inp | Input file (model definition) |
| .lck | Lock file (exists while job runs) |
Troubleshooting
| Status/Error | Meaning | Solution |
|---|---|---|
| COMPLETED | Success | Proceed to /abaqus-odb |
| ABORTED | Failed | Check .msg file for error |
| License not available | No tokens | Wait or check license server |
| Memory error | Model too large | Increase memory or coarsen mesh |
| .lck file exists | Stale lock | Delete if job is not running |
Validation Checklist
Before submitting:
- Model saved (.cae exists)
- Job name specified
- Model name matches saved model
- CPUs set appropriately
Code Patterns
For API syntax and code examples, see:
Version History
- 2178ed3 Current 2026-07-24 17:39
