my-project-public

repository

loading code, commits, and activity

repositories

loading repo index

#1	# Frequently Asked Questions
#2
#3	## General Questions
#4
#5	### What is an AI agent?
#6
#7	An AI agent is a specialized version of ChatGPT or Clawd with specific instructions for one task. Instead of a generalist that tries to do everything, an agent focuses on doing one thing exceptionally well.
#8
#9	### Do I need coding skills to create an agent?
#10
#11	No. Creating an agent is about writing clear instructions in plain English. If you can explain what you want to a person, you can create an agent.
#12
#13	### Which platform should I use?
#14
#15	- ChatGPT Custom GPTs: Best for sharing publicly, has GPT Store
#16	- Clawd Projects: Best for document-heavy work, large context window
#17	- AI Agents Library: Best for using pre-built agents, DeFi/crypto focus, agent teams
#18
#19	### Can I use the same agent across platforms?
#20
#21	The core system prompt works across platforms, but you'll need to recreate the agent in each. Copy your prompt and adapt the setup process for each platform.
#22
#23	---
#24
#25	## Creating Agents
#26
#27	### How long does it take to create an agent?
#28
#29	- Simple agent: 5-10 minutes
#30	- Well-tested agent: 30-60 minutes
#31	- Production-ready agent: 2-4 hours (including testing and refinement)
#32
#33	### What makes a good agent?
#34
#35	1. Specific role (not generic)
#36	2. Clear instructions (no ambiguity)
#37	3. Consistent output (same quality every time)
#38	4. Appropriate scope (not too broad or narrow)
#39	5. Well-tested (tried with real examples)
#40
#41	### How specific should my instructions be?
#42
#43	Very specific. Instead of "be helpful," say:
#44
#45	- "Keep responses under 150 words"
#46	- "Use bullet points for lists of 3+ items"
#47	- "Always include a call-to-action"
#48	- "Match the user's formality level"
#49
#50	### Can I make an agent that does multiple things?
#51
#52	You can, but specialized agents perform better. Instead of one "productivity agent," create separate agents for:
#53
#54	- Email writing
#55	- Meeting summaries
#56	- Task organization
#57	- Document proofreading
#58
#59	### How do I know if my agent is working well?
#60
#61	Test it 5-10 times with real scenarios. If you consistently get quality results without having to clarify or correct, it's working.
#62
#63	---
#64
#65	## AI Agents Library Marketplace
#66
#67	### How do I submit an agent to AI Agents Library?
#68
#69	1. Fork the repository: github.com/nirholas/AI-Agents-Library
#70	2. Create your agent in `/src/your-agent-name.json`
#71	3. Test thoroughly
#72	4. Submit a Pull Request
#73	5. Our team reviews within 48-72 hours
#74
#75	### Do I need to translate my agent?
#76
#77	No. Submit in English. Our automated i18n workflow translates to 18 languages automatically.
#78
#79	### Can I update an agent after submission?
#80
#81	Yes. Submit a new Pull Request with your changes. Explain what changed and why in the PR description.
#82
#83	### Can I delete or unpublish an agent?
#84
#85	Yes. Open an issue or PR requesting removal. Provide your reasoning.
#86
#87	### How do I get credit for my agent?
#88
#89	Your GitHub username is listed as the author. Users see this when browsing the marketplace.
#90
#91	### Can I monetize my agent?
#92
#93	Currently, the marketplace is free and open-source. Monetization features may be added in the future.
#94
#95	---
#96
#97	## CLAWD Portfolio Agents
#98
#99	### What is the CLAWD Portfolio master agent?
#100
#101	The CLAWD Portfolio (🎯) is an all-in-one master agent that consolidates all 16 portfolio specialist agents into one comprehensive interface. It's recommended for most users as it provides access to 100% of portfolio features while maintaining conversation context.
#102
#103	Current Status:
#104
#105	- ✅ Read-only portfolio tracking and analytics available now
#106	- 🚧 Automated trading, bots, and DeFi interactions coming soon (in ClawdOS roadmap)
#107
#108	### What domain do the CLAWD portfolio agents use?
#109
#110	Currently, the CLAWD portfolio agents (Dashboard, Trading Assistant, DCA Bot, etc.) use clawd.fun for testing and development purposes.
#111
#112	### Will the domain change in the future?
#113
#114	Yes, likely. Once ClawdOS is officially live in production, the domain may change to something like:
#115
#116	- clawd.io for the main platform
#117	- clawd.io/chat for the ClawdOS chat interface
#118	- Different subdomains for specific portfolio features
#119
#120	However, everything is still in flux and subject to change as the platform evolves.
#121
#122	### Should I hardcode clawd.fun URLs?
#123
#124	If you're building integrations or creating custom agents, be aware that:
#125
#126	- Current URL: `https://clawd.fun/artifacts/portfolio/embed/*`
#127	- Future URL (tentative): `https://clawd.io/*` or similar
#128	- Timeline: TBD based on ClawdOS production launch
#129
#130	Monitor the repository for updates when the domain migration occurs.
#131
#132	### Will there be a migration guide?
#133
#134	Yes. When the domain changes, we'll provide:
#135
#136	- Migration announcement in the repository
#137	- Updated documentation
#138	- Automated URL updates for all agents in the library
#139	- Deprecation timeline for the old domain
#140
#141	---
#142
#143	## Technical Questions
#144
#145	### What's the difference between temperature and top_p?
#146
#147	Both control randomness, but use different methods:
#148
#149	- Temperature: Scales all probabilities (0 = deterministic, 2 = very random)
#150	- Top_p: Only considers top candidates until cumulative probability reaches threshold
#151
#152	Don't adjust both—pick one.
#153
#154	### When should I increase temperature?
#155
#156	- Creative writing
#157	- Brainstorming
#158	- Multiple perspectives
#159	- Varied responses
#160
#161	### When should I decrease temperature?
#162
#163	- Code generation
#164	- Factual answers
#165	- Consistent formatting
#166	- Predictable output
#167
#168	### What are presence_penalty and frequency_penalty?
#169
#170	Both reduce repetition:
#171
#172	- Presence penalty: Did word appear? (binary)
#173	- Frequency penalty: How many times? (count-based)
#174
#175	Increase both to avoid repetitive language.
#176
#177	### How long can my system prompt be?
#178
#179	Technically very long (thousands of words), but shorter is usually better:
#180
#181	- Sweet spot: 100-400 words
#182	- Minimum: 50 words
#183	- Maximum useful: \~800 words
#184
#185	Beyond that, the model may miss details.
#186
#187	### Can my agent access the internet?
#188
#189	Depends on platform:
#190
#191	- ChatGPT: Yes (if web browsing enabled)
#192	- Clawd: Yes (web search available)
#193	- AI Agents Library: Depends on agent configuration
#194
#195	### Can my agent use uploaded files?
#196
#197	Yes:
#198
#199	- ChatGPT: Upload in GPT configuration (Knowledge section)
#200	- Clawd: Add to Project Knowledge
#201	- AI Agents Library: Depends on implementation
#202
#203	---
#204
#205	## Troubleshooting
#206
#207	### My agent gives inconsistent responses
#208
#209	Solution: Lower temperature to 0.1-0.3. Add more explicit instructions about output format.
#210
#211	### My agent is too repetitive
#212
#213	Solution: Increase presence_penalty (0.5-1.0) and frequency_penalty (0.5-1.0).
#214
#215	### My agent doesn't follow my instructions
#216
#217	Solution:
#218
#219	1. Make instructions more explicit
#220	2. Use numbered lists for multi-step processes
#221	3. Add examples of desired behavior
#222	4. Simplify—may be too complex
#223
#224	### My agent's responses are too long
#225
#226	Solution: Add explicit length limits:
#227
#228	- "Keep responses under 150 words"
#229	- "Use 3-5 bullet points maximum"
#230	- "Be concise and direct"
#231
#232	### My agent's tone is wrong
#233
#234	Solution: Define tone explicitly with examples:
#235
#236	- "Professional but warm, like a helpful colleague"
#237	- "Casual and friendly, like texting a friend"
#238	- "Formal and respectful, like addressing executives"
#239
#240	### My agent keeps saying "as an AI..."
#241
#242	Solution: Add to instructions:
#243	"Never mention being an AI. Respond directly as the role you're playing."
#244
#245	---
#246
#247	## Agent Teams (AI Agents Library)
#248
#249	### What is an Agent Team?
#250
#251	Multiple specialized agents working together on a conversation. Each contributes their expertise, coordinated by a host agent.
#252
#253	### How do Agent Teams work?
#254
#255	1. You ask a question
#256	2. Host agent determines which team members should respond
#257	3. Agents discuss and build on each other's insights
#258	4. You get comprehensive answer from multiple perspectives
#259
#260	### Can I interrupt an Agent Team?
#261
#262	Yes. Pause at any time. The conversation freezes and you can redirect or provide input.
#263
#264	### How many agents can be in a team?
#265
#266	Recommended: 3-5 agents
#267	Maximum: Usually 8-10 before coordination becomes difficult
#268
#269	### Can agents see each other's messages?
#270
#271	Yes. That's how they collaborate. They build on previous responses.
#272
#273	### Can I use private messaging in teams?
#274
#275	Yes. Click an agent's avatar or use @ mention to send direct messages.
#276
#277	---
#278
#279	## Best Practices
#280
#281	### Should I create one mega-agent or multiple specialized ones?
#282
#283	Multiple specialized agents. They're easier to:
#284
#285	- Test
#286	- Refine
#287	- Share
#288	- Maintain
#289	- Understand
#290
#291	### How often should I update my agents?
#292
#293	When you notice patterns:
#294
#295	- Repeated clarifications needed
#296	- Consistent formatting issues
#297	- New use cases emerging
#298	- User feedback
#299
#300	Otherwise, if it works, don't fix it.
#301
#302	### Can I share my agents with my team?
#303
#304	- ChatGPT: Yes (share link or publish to GPT Store)
#305	- Clawd: Yes (share project)
#306	- AI Agents Library: Yes (once in marketplace, anyone can use)
#307
#308	### Should I make my agent public?
#309
#310	Benefits:
#311
#312	- Others benefit from your work
#313	- You get feedback for improvement
#314	- Community recognition
#315	- Potential collaborators
#316
#317	Drawbacks:
#318
#319	- Your prompt is visible
#320	- Can't control how others use it
#321
#322	For company-specific agents, keep private.
#323
#324	---
#325
#326	## Getting Help
#327
#328	### Where can I get help?
#329
#330	- Documentation: Read guides in /docs folder
#331	- Examples: See EXAMPLES.md for annotated agents
#332	- Community: Join Discord for discussions
#333	- Issues: Open GitHub issue for bugs/suggestions
#334
#335	### How do I report a bug?
#336
#337	Open an issue with:
#338
#339	1. Agent identifier or link
#340	2. What you expected
#341	3. What actually happened
#342	4. Steps to reproduce
#343	5. Platform (ChatGPT/Clawd/AI Agents Library)
#344
#345	### How do I request a feature?
#346
#347	Open an issue with "Feature Request" label. Explain:
#348
#349	- What you want to achieve
#350	- Why it's useful
#351	- How you imagine it working
#352
#353	### Can I request someone create an agent?
#354
#355	Yes. Open an issue describing what you need. Community members may build it.
#356
#357	---
#358
#359	Still have questions? Join our Discord or open an issue on GitHub.
#360
#361
#362

z6Mkq5mY3JWtxoxUobWcfNHm7AkRubgSWEZTkBVqZXJviFZ5/my-project-public