> ## Documentation Index
> Fetch the complete documentation index at: https://aspex.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# aspex simulate

> Counterfactual security analysis: what would happen to your attack surface if you changed something. Nothing is modified.

```sh theme={"dark"}
aspex simulate --remove-server playwright
aspex simulate --restrict-filesystem filesystem=~/projects/acme
aspex simulate --deny-network '*'
aspex simulate --remove-tool desktop-commander.start_process
```

`simulate` clones your environment in memory, applies a hypothetical change, recomputes capabilities and attack paths through the same pipeline as a real scan, and compares before with after. Your configuration is never touched; the last line of every run says so.

## What you can change

| Flag                                         | Effect                                                                                   |
| -------------------------------------------- | ---------------------------------------------------------------------------------------- |
| `--remove-server NAME`                       | drop a configured server                                                                 |
| `--restrict-filesystem [SERVER=]ROOT[,ROOT]` | replace a filesystem server's allowed roots (omit `SERVER=` for every filesystem server) |
| `--deny-network [SERVER\|*]`                 | remove network egress, as an allowlist of nothing would                                  |
| `--remove-tool SERVER.TOOL`                  | drop one tool from a live tool list                                                      |
| `--remove-hook EVENT`                        | remove lifecycle hooks for an event                                                      |
| `--remove-skill NAME`                        | remove a skill                                                                           |

Flags are repeatable and combine, so you can evaluate a whole hardening change at once.

## Output

```
$ aspex simulate --restrict-filesystem filesystem=~/projects/acme

  ◆  Security impact simulation
     • restrict filesystem to ~/projects/acme

  BEFORE   blast radius HIGH   6 attack path(s)
  AFTER    blast radius MEDIUM 2 attack path(s)

  REMOVED ATTACK PATHS
     ✓ Potential sensitive data exfiltration path  CRITICAL · filesystem + playwright
     ✓ Potential persistent agent compromise path  CRITICAL · filesystem + brave-search

  SEVERITY CHANGED
     ~ Potential sensitive data exfiltration path (filesystem + github): HIGH → MEDIUM

  Capability changes
     ~ filesystem  roots /Users/you → /Users/you/projects/acme

  No configuration was modified.
```

`--json` emits the versioned `aspex-simulate/v1` schema: before/after environments, capability deltas, and attack paths added and removed.

## Where simulation shows up elsewhere

* **`aspex explain "…"`** ends every YES answer with the controls that break the path, each simulated so you see which one removes it on its own.
* **`aspex tighten`** runs each recommendation through the simulator and shows its security impact next to the functional impact.
* **`aspex inspect <command|path>`** simulates adding a server you have not installed yet.
* **`aspex mcp`** exposes `aspex_simulate_change` so your coding agent can check a proposed `.mcp.json` before it writes it.

All four share this one engine, so the numbers agree everywhere.