Build, organize, and maintain a library of tested prompts for marketing, coding, and operations. Use when saving effective prompts, creating prompt templates, organizing prompts by category, or building a reusable prompt collection. Triggers on "save this prompt", "prompt library", "prompt template", "add to prompts", or any prompt management request.
HONEST CLASSIFICATION: This is a prompt template, not an executable plugin. OpenClaw injects this as context to guide LLM behavior. It does NOT enforce rules at runtime. Real enforcement requires the shell scripts in
OpenClawData/scripts/.
Build and maintain an organized library of tested, reusable prompts.
qwen3:8b — for crafting and refining prompt templates.
OpenClawData/prompts/OpenClawData/memory/prompt-templates.mdOrganize prompts into category files:
| File | Category | Examples |
|---|
marketing-prompts.md | Marketing & content | Caption generation, campaign ideas, repurposing |
coding-prompts.md | Coding & automation | Script templates, debugging, API integration |
planning-prompts.md | Strategy & planning | Task breakdown, campaign planning, research |
operations-prompts.md | Operations & reporting | Reports, checklists, summaries |
### <Prompt Name>
**Category:** <category>
**Model:** qwen3:8b | qwen2.5-coder:7b
**Tested:** yes | no
**Rating:** ★☆☆☆☆ to ★★★★★
**Template:**
<the actual prompt with {placeholders} for variable parts>
**Variables:**
- `{variable1}` — <description>
- `{variable2}` — <description>
**Example usage:**
Notes: <any tips, caveats, or variations>
## Library Rules
1. **Only save tested prompts** — a prompt must produce good output at least once before being added
2. **Include the model** — which model was this tested on?
3. **Rate honestly** — ★★★ means "works but could be better"
4. **Use placeholders** — make prompts reusable with `{variables}`
5. **Organize by category** — one file per category, not one giant file
6. **Update, don't duplicate** — if a prompt improves, update the existing entry
## Adding a New Prompt
1. Test the prompt with the appropriate model
2. If output is good, save using the template format above
3. Append to the correct category file in `OpenClawData/prompts/`
4. Also add a brief reference in `OpenClawData/memory/prompt-templates.md`