====== OpenSpec ======

https://dev.to/webdeveloperhyper/how-to-make-ai-follow-your-instructions-more-for-free-openspec-2c85

===== install =====

<code bash>
yay -S openspec
</code>

===== add openspec to existing project =====

cd <project>
<code bash>
openspec init
</code>

from agent auto compile **openspec\project.md** with command
<code bash>
Please read openspec/project.md and help me fill out with details about my project, tech stack, and conventions.
</code>

===== add new feature =====

==== Proposal ====


<code bash>
/openspec:proposal load multiple svm models
</code>

file created
<code>
openspec/changes/
└── load-multiple-svm-models
    ├── proposal.md
    ├── specs
    │   └── svm-anomaly-detection
    │       └── spec.md
    └── tasks.md
</code>


Proposal Structure:

  * **proposal.md**: Explains why multiple models are needed (flexibility for different scenarios), what changes will be made, and the impact
  * **tasks.md**: Detailed implementation checklist broken into 5 phases (design, configuration, processor modification, main function updates, and testing)
  * **spec.md**: Specification deltas with both MODIFIED (existing SVM loading requirement) and ADDED (new model configuration requirement) sections, including proper scenarios

eventually check status of proposals
<code bash>
openspec list    # Confirm the change folder exists
openspec validate load-multiple-svm-models
openspec show load-multiple-svm-models   # Review proposal, tasks, and spec delta
</code>


==== Apply ====

coding
<code bash>
/openspec:apply load-multiple-svm-models
</code>

...testing...

==== Archive ====

After implementation and testing, archive the change. All of the checkboxes in the TODO list were checked, and files were moved to the archive folder.
<code>
#openspec archive load-multiple-svm-models --yes
openspec archive load-multiple-svm-models --skip-specs --yes
</code>