| name | shesha-skill-creator |
| description | Creates new Shesha-specific skills by analyzing existing Shesha application code, documentation, and descriptions. Produces properly structured SKILL.md files with supporting reference files following best practices. Use when the user wants to create a new skill for Shesha code generation, workflow patterns, or any Shesha-specific domain by referencing existing implementations. |
Shesha Skill Creator
Create new skills for Shesha applications by analyzing existing code and documentation provided by the user via $ARGUMENTS.
For general skill authoring best practices, consult ../skill-creator/SKILL.md and its references/ directory.
Workflow
1. Gather Inputs
Accept one or more of:
- Code references: Paths to existing Shesha project directories or files (e.g.
backend/src/Module/SaGov.Leave.Application/)
- Documentation: URLs or file paths to Shesha-related docs
- Descriptions: Natural language description of what the skill should generate
If ambiguous, ask:
- "What artifacts should this skill generate?"
- "Which existing code best represents the patterns to follow?"
- "What would a user say that should trigger this skill?"
2. Analyze Referenced Code
Read the referenced code and extract patterns using this Shesha-specific checklist:
Base classes and inheritance
Attributes and decorators
DI and infrastructure patterns
Validation patterns
NHibernate requirements
Folder conventions
Record for each artifact type found:
- File name pattern
- Location/folder convention
- Full code template (with
{Placeholder} tokens)
- Required using statements
- Key rules and constraints
3. Plan the Skill Structure
Based on analysis, decide:
Naming: Use kebab-case describing the domain (e.g. shesha-leave-management, shesha-reporting).
Artifact catalog: List all artifact types the skill will generate, grouped logically.
File split strategy (target: SKILL.md under 500 lines):
- SKILL.md: Overview, artifact catalog table, folder structure, quick reference tables
- Supporting files: Full code templates grouped by layer or domain area
- Use
§N section numbering in supporting files
Reference the existing skills as structural models:
4. Create the Skill
Create .claude/skills/{skill-name}/ with:
SKILL.md — Follow this structure:
---
name: {skill-name}
description: {Third-person description of what it generates and when to use it. Include trigger words.}
---
# {Skill Title}
Generate {artifact types} for a Shesha application based on $ARGUMENTS.
## Instructions
- {Key rules specific to this domain}
## Artifact catalog
| # | Artifact | Layer | Template |
|---|----------|-------|----------|
| 1 | {Name} | {Domain/Application} | [{file}.md]({file}.md) §1 |
## Folder structure
{Expected file tree}
## Quick reference
{Base classes, attributes, common patterns tables}
Now generate the requested artifact(s) based on: $ARGUMENTS
Supporting files — One per logical group, using §N sections:
# {Group Title}
## §1. {Artifact Name}
**File:** `{FileName}` in `{Location}/`
{Code template with {Placeholder} tokens}
**Key rules:**
- {Constraint 1}
- {Constraint 2}
5. Verify
Now create the requested skill based on: $ARGUMENTS