MEGA-Code Help

Available Commands

Command	Description
/mega-code:login	Sign in via GitHub or Google OAuth
/mega-code:run	Run skill extraction pipeline
/mega-code:status	Show pending items and status
/mega-code:help	Show this help

Output Locations

Type	Pending Location	Installed Location
Skills	~/.local/share/mega-code/data/pending-skills/{name}/	.claude/skills/{name}/SKILL.md
Strategies	~/.local/share/mega-code/data/pending-strategies/{name}.md	.claude/rules/mega-code/{name}.md

Skill Structure

Generated skills follow this structure:

~/.local/share/mega-code/data/pending-skills/{skill-name}/
├── SKILL.md        # Main skill content
├── injection.json  # Auto-trigger rules
├── evidence.json   # Source evidence
└── metadata.json   # Generation info

Strategy Structure

Strategies are modular rules saved as:

---
paths: **/*.py
---

# Strategy Title

Clear statement of the preference or convention.

Model Options

The --model flag for /mega-code:run accepts any model alias supported by the LLM module:

Alias	Provider
gemini-3-flash	Google
gpt-5-mini	OpenAI

When no model is specified, the server selects the best model based on your configured LLM keys.

Tips

• Run /mega-code:run after significant coding sessions
• Use --project to analyze multiple sessions for stronger patterns
• Skills with more evidence (from multiple sessions) are higher quality
• Review and edit skills before installing for best results
• For plugin updates, use /plugin marketplace update mind-ai-mega-code