HippocampusEvolve/universal-agent-seed
Self-improving, stack-agnostic agentic starter: AGENTS.md + mise, a curated MCP toolset (Context7/Serena/Playwright/GitHub), and a knowledge/skills/hooks growth loop.
stack: Python
ARCHETYPE
Fort Knox
Deny lists, pre-tool hooks, sandboxing. Nothing touches prod without a signature.
CONTEXT TAX · EVERY TURN
up to ~37.2k tokens
Claude Code loads MCP tools on demand, so the real cost is usually lower.
Context hog · median rig: 2.3k · breakdown
GUARDRAILS
5/5
Blocks destructive commands · Protects secrets · Pre-tool screening hook · No YOLO mode · Sandbox or ask-first rules · details
Quality 76/100 Guardrails v2
Copy this rig
# review before running: this installs third-party code $ claude mcp add --transport http context7 https://mcp.context7.com/mcp $ claude mcp add serena -- uvx --from git+https://github.com/oraios/serena serena start-mcp-server --context ide-assistant --project . $ claude mcp add playwright -- npx -y @playwright/mcp@latest $ claude mcp add --transport http github https://api.githubcopilot.com/mcp/ -H 'Authorization: YOUR_VALUE' $ npx degit HippocampusEvolve/universal-agent-seed/.claude ./rig-universal-agent-seed # inspect, then merge into .claude/
MCP servers are added to Claude Code at local scope; env vars are shown as YOUR_… placeholders — we never store values. Files are fetched with degit into a separate folder so you can review before merging.
$ claude mcp add --transport http context7 https://mcp.context7.com/mcp $ claude mcp add serena -- uvx --from git+https://github.com/oraios/serena serena start-mcp-server --context ide-assistant --project . $ claude mcp add playwright -- npx -y @playwright/mcp@latest $ claude mcp add --transport http github https://api.githubcopilot.com/mcp/ -H 'Authorization: YOUR_VALUE'
$ npx degit HippocampusEvolve/universal-agent-seed/.claude/skills/bootstrap-stack .claude/skills/bootstrap-stack $ npx degit HippocampusEvolve/universal-agent-seed/.claude/skills/distill .claude/skills/distill $ npx degit HippocampusEvolve/universal-agent-seed/.claude/skills/extract-version .claude/skills/extract-version $ npx degit HippocampusEvolve/universal-agent-seed/.claude/skills/learn .claude/skills/learn $ npx degit HippocampusEvolve/universal-agent-seed/.claude/skills/reflect .claude/skills/reflect $ npx degit HippocampusEvolve/universal-agent-seed/.claude/skills/seed-update .claude/skills/seed-update
$ curl -fsSL --create-dirs -o .claude/agents/code-reviewer.md https://raw.githubusercontent.com/HippocampusEvolve/universal-agent-seed/main/.claude/agents/code-reviewer.md $ curl -fsSL --create-dirs -o .claude/agents/researcher.md https://raw.githubusercontent.com/HippocampusEvolve/universal-agent-seed/main/.claude/agents/researcher.md $ curl -fsSL --create-dirs -o .claude/agents/test-writer.md https://raw.githubusercontent.com/HippocampusEvolve/universal-agent-seed/main/.claude/agents/test-writer.md
Merge into .claude/settings.json (project) or ~/.claude/settings.json (user). Hook commands reference scripts in the source repo — copy those too.
{
"permissions": {
"deny": [
"Read(.env)",
"Read(**/*.pem)",
"Read(**/*.key)",
"Read(**/*.p12)",
"Read(**/*.pfx)",
"Read(**/id_rsa*)",
"Read(**/id_ed25519*)",
"Read(**/credentials.json)",
"Read(**/credentials.yml)",
"Read(**/credentials.yaml)",
"Read(secrets/**)"
],
"ask": [
"Bash(git push:*)",
"Bash(rm:*)"
]
},
"hooks": {
"PreToolUse": [
{
"matcher": "Bash|Edit|Write|MultiEdit|NotebookEdit|Read",
"hooks": [
{
"type": "command",
"command": "python \"${CLAUDE_PROJECT_DIR}/.claude/hooks/protect_files.py\""
}
]
}
]
}
} MCP servers (4)
| server | source | est. tokens |
|---|---|---|
| Context7 | remote · remote | 1.2k |
| Serena | git | 8.0k |
| Playwright MCP · "playwright" | npm | 7.5k |
| GitHub MCP · "github" | remote · remote | 18.0k |
Skills (6)
Subagents (3)
| code-reviewer model: inherit | >- |
| researcher model: sonnet | >- |
| test-writer | >- |
Hooks (4)
| event | matcher | runs |
|---|---|---|
| SessionStart | * | python "${CLAUDE_PROJECT_DIR}/.claude/hooks/inject_knowledge.py" |
| SessionEnd | * | python "${CLAUDE_PROJECT_DIR}/.claude/hooks/capture_learnings.py" |
| PreToolUse | Bash|Edit|Write|MultiEdit|NotebookEdit|Read | python "${CLAUDE_PROJECT_DIR}/.claude/hooks/protect_files.py" |
| PostToolUse | Edit|Write|MultiEdit | python "${CLAUDE_PROJECT_DIR}/.claude/hooks/format_after_edit.py" |
Cursor rules (1)
00-defer-to-agents-md · always
Permissions
deny (11)
Read(.env)
Read(**/*.pem)
Read(**/*.key)
Read(**/*.p12)
Read(**/*.pfx)
Read(**/id_rsa*)
Read(**/id_ed25519*)
Read(**/credentials.json)
Read(**/credentials.yml)
Read(**/credentials.yaml)
Read(secrets/**)
ask (2)
Bash(git push:*)
Bash(rm:*)
allow (13)
Bash(mise:*)
Bash(dprint:*)
Bash(lefthook:*)
Bash(ls:*)
Bash(git status:*)
Bash(git diff:*)
Bash(git log:*)
Bash(git show:*)
Bash(git branch:*)
Bash(git add:*)
Bash(git commit:*)
Read(.env.example)
Read(.env.sample)
Similar rigs
Lamtea/dotfiles
My dotfiles for command-line interface on Arch linux.
Pragmatist 37.5k tok · · Q0
codeverbojan/claude-code-kickstart
Production-grade agentic workflow template for Claude Code. 4-layer architecture, task playbooks, session memory, sub-agents, one-click install.
Orchestrator 10.2k tok · · Q15
BitWise-0x/.dot-files
My .dot-files
Pragmatist 27.6k tok · · Q0
okadrive/dotfiles
—
Pragmatist 27.6k tok · · Q0