Skip to content
Back to skills

Tailscale Mesh

ASecurity

Tailscale mesh VPN for secure peer-to-peer networking. WireGuard-based overlay network with MagicDNS and ACLs.

  • 61 stars
  • 0 votes
  • 0 copies
  • 2 views
  • Added May 29, 2026
developmentpythongobash

Works with

  • mcp

Security analysis

A100/100

Scanned May 29, 2026

npx -y skills add plurigrid/asi --skill tailscale-mesh --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Tailscale Mesh?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Tailscale Mesh
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/plurigrid-tailscale-mesh/badge)](https://www.skillsdirectory.com/skills/plurigrid-tailscale-mesh)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: tailscale-mesh
description: Tailscale mesh VPN for secure peer-to-peer networking. WireGuard-based overlay network with MagicDNS and ACLs.
version: 1.0.0
---


# Tailscale Mesh Skill

**Trit**: 0 (ERGODIC - mediates network topology)  
**Foundation**: Tailscale + WireGuard + DERP  

## Core Concept

Tailscale creates a mesh VPN:
- WireGuard encryption
- NAT traversal via DERP relays
- MagicDNS for hostname resolution
- ACLs for access control

## Common Commands

```bash
# Status
tailscale status
tailscale netcheck

# Connect/disconnect
tailscale up
tailscale down

# Send files
tailscale file cp file.txt hostname:

# SSH
tailscale ssh hostname

# Funnel (public exposure)
tailscale funnel 8080
```

## ACL Configuration

```jsonc
{
  "acls": [
    {"action": "accept", "src": ["group:dev"], "dst": ["*:*"]},
    {"action": "accept", "src": ["tag:server"], "dst": ["tag:db:5432"]}
  ],
  "tagOwners": {
    "tag:server": ["group:ops"],
    "tag:db": ["group:dba"]
  }
}
```

## GF(3) Integration

```python
def trit_from_connection(conn):
    """Map connection type to GF(3) trit."""
    if conn.type == "direct":
        return 1   # PLUS: optimal path
    elif conn.type == "derp":
        return 0   # ERGODIC: relayed
    else:
        return -1  # MINUS: failed/blocked
```

## Canonical Triads

```
bisimulation-game (-1) ⊗ tailscale-mesh (0) ⊗ localsend-mcp (+1) = 0 ✓
spi-parallel-verify (-1) ⊗ tailscale-mesh (0) ⊗ tailscale-file-transfer (+1) = 0 ✓
```

## See Also

- `tailscale-file-transfer` - File transfer with open games semantics
- `localsend-mcp` - P2P transfer via LocalSend



## Scientific Skill Interleaving

This skill connects to the K-Dense-AI/claude-scientific-skills ecosystem:

### Graph Theory
- **networkx** [○] via bicomodule
  - Universal graph hub

### Bibliography References

- `graph-theory`: 38 citations in bib.duckdb
- `distributed-systems`: 3 citations in bib.duckdb

## Cat# Integration

This skill maps to **Cat# = Comod(P)** as a bicomodule in the equipment structure:

```
Trit: 0 (ERGODIC)
Home: Prof
Poly Op: ⊗
Kan Role: Adj
Color: #26D826
```

### GF(3) Naturality

The skill participates in triads satisfying:
```
(-1) + (0) + (+1) ≡ 0 (mod 3)
```

This ensures compositional coherence in the Cat# equipment structure.

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.