my-project-public

repository

loading code, commits, and activity

repositories

loading repo index

my-project-public — gitlawb

#1	---
#2	name: pump-ai-agents
#3	description: "AI agent integration layer for the Pump SDK — agent instruction files, .well-known discovery, LLM context documents, 15+ skill files, MCP server prompts, and terminal management rules for GitHub Copilot, Claude, and Gemini."
#4	metadata:
#5	openclaw:
#6	homepage: https://github.com/nirholas/pump-fun-sdk
#7	---
#8
#9	# AI Agent Integration — Agent Scaffolding & Discovery
#10
#11	Configure AI agents (GitHub Copilot, Claude, Gemini) to work effectively with the Pump SDK through instruction files, skill registries, and MCP server integration.
#12
#13	## Agent Instruction Files
#14
#15	\| File \| Agent \| Purpose \|
#16	\|------\|-------\|---------\|
#17	\| `AGENTS.md` \| Universal \| Project overview, architecture, security rules \|
#18	\| `CLAUDE.md` \| Claude Code \| Key patterns, critical rules \|
#19	\| `COPILOT.md` \| GitHub Copilot \| Instructions pointer \|
#20	\| `GEMINI.md` \| Google Gemini \| Instructions pointer \|
#21	\| `.github/copilot-instructions.md` \| Copilot Chat \| SDK patterns, security rules \|
#22
#23	## .well-known Discovery Files
#24
#25	\| File \| Purpose \|
#26	\|------\|---------\|
#27	\| `.well-known/ai-plugin.json` \| AI plugin manifest \|
#28	\| `.well-known/agent.json` \| Agent capabilities and config \|
#29	\| `.well-known/skills.json` \| Skills registry (15+ skills) \|
#30	\| `.well-known/security.txt` \| Security contact information \|
#31
#32	## LLM Context Documents
#33
#34	\| File \| Size \| Purpose \|
#35	\|------\|------\|---------\|
#36	\| `llms.txt` \| Short \| Quick reference for LLMs \|
#37	\| `llms-full.txt` \| Long \| Comprehensive LLM context \|
#38
#39	## Skills Registry Schema
#40
#41	```json
#42	{
#43	"schema_version": "v1",
#44	"name": "pump-fun-sdk",
#45	"skills": [
#46	{
#47	"id": "pump-sdk-core",
#48	"name": "Pump SDK Core",
#49	"description": "...",
#50	"path": ".github/skills/pump-sdk-core.skill.md",
#51	"tags": ["typescript", "solana", "sdk"]
#52	}
#53	]
#54	}
#55	```
#56
#57	## Terminal Management (Agent Rules)
#58
#59	- Always use background terminals (`isBackground: true`) for every command
#60	- Always kill the terminal after completion — never leave terminals open
#61	- Do not reuse foreground shell sessions
#62	- If a terminal appears unresponsive, kill it and create a new one
#63
#64	## Security Invariants for Agents
#65
#66	- ONLY official Solana Labs crypto: `solana-sdk`, `@solana/web3.js`, `solana-keygen`
#67	- Zeroize all key material after use
#68	- File permissions `0600` for keypairs
#69	- No network calls for key generation
#70	- All amounts use `BN` (bn.js) — never JavaScript `number` for financial math
#71	- `createInstruction` (v1) is deprecated — use `createV2Instruction`
#72
#73	## MCP Server Integration
#74
#75	```json
#76	{
#77	"mcpServers": {
#78	"solana-wallet": {
#79	"command": "npx",
#80	"args": ["-y", "@pump-fun/mcp-server"],
#81	"transportType": "stdio"
#82	}
#83	}
#84	}
#85	```
#86
#87	## Patterns to Follow
#88
#89	- Keep agent instruction files concise and actionable
#90	- Point agents to specific skill files for domain knowledge
#91	- Update `skills.json` when adding or modifying skill files
#92	- Maintain consistency between `AGENTS.md`, `CLAUDE.md`, and `COPILOT.md`
#93
#94	## Common Pitfalls
#95
#96	- Agent instruction files that are too long get truncated or ignored
#97	- Missing skills from `skills.json` means agents won't discover them
#98	- Terminal management rules are critical in Codespaces — stale terminals block operations
#99	- MCP server must be registered in the agent's tool configuration
#100
#101

z6Mkq5mY3JWtxoxUobWcfNHm7AkRubgSWEZTkBVqZXJviFZ5/my-project-public