Purpose
Establish a secure, verified path when access is blocked by geo/IP policy, then resume the blocked workflow safely and audibly.
Primary outcomes:
- detect and classify block behavior,
- switch to a valid tunnel path with explicit user consent,
- verify public IP, region, and DNS safety posture,
- re-run blocked task with bounded retries,
- return an auditable connection report.
This is an orchestration skill. It does not guarantee legal access to restricted services.
Required Installed Skills
Core diagnostics/orchestration:
shell-scripting (inspected latest: 1.0.0)
curl-http (inspected latest: 1.0.0)
Tunnel path options (at least one):
- provider CLI path (NordVPN / Mullvad / ExpressVPN) via shell orchestration
wireguard (inspected latest: 1.0.0)
tailscale (inspected latest: 1.0.0)
Safety and verification extensions:
dns (inspected latest: 1.0.0)
ipinfo (inspected latest: 1.0.0)
moltguard (inspected latest: 6.0.2, optional but recommended)
Install/update:
npx -y clawhub@latest install shell-scripting
npx -y clawhub@latest install curl-http
npx -y clawhub@latest install wireguard
npx -y clawhub@latest install tailscale
npx -y clawhub@latest install dns
npx -y clawhub@latest install ipinfo
npx -y clawhub@latest install moltguard
npx -y clawhub@latest update --all
Verify:
npx -y clawhub@latest list
Required Credentials and Access
Required access:
- valid account/session for selected tunnel path
- local executable for selected path (
nordvpn/mullvad/expressvpn or wg or tailscale)
Optional keys:
MOLTGUARD_API_KEY (if MoltGuard remote detection mode is enabled)
IPINFO_TOKEN (optional, higher quota geolocation verification)
Preflight:
command -v nordvpn || command -v mullvad || command -v expressvpn || command -v wg || command -v tailscale
echo "$MOLTGUARD_API_KEY" | wc -c
echo "$IPINFO_TOKEN" | wc -c
Mandatory behavior:
- Never fail silently on missing keys/auth.
- Always return
MissingAPIKeys and/or MissingCredentials with blocked stages.
- Continue with non-blocked diagnostics and mark output as
Partial when needed.
Compliance Gate (Mandatory)
Before any tunnel switch, confirm and record:
- user authorization to modify network routing,
- acknowledgment of legal/terms responsibility,
- stated purpose for geo-switch (testing, parity checks, privacy hardening).
If acknowledgment is missing:
- do not execute switching commands,
- return diagnostics-only output.
Inputs the LM Must Collect First
blocked_url or blocked_endpoint
blocked_task_name (example: prediction-market-arbitrage)
target_region
tunnel_path (provider-cli, wireguard, tailscale-exit-node)
provider_or_profile (provider name, WG profile, or exit-node name)
risk_mode (diagnose-only, switch-and-verify, switch-and-resume)
kill_switch_required (yes/no)
max_retries (default: 2)
Do not execute switching before tunnel path and target region are explicit.
Tool Responsibilities
shell-scripting
Use as control plane:
- executable detection,
- connect/disconnect wrappers,
- retry and cleanup logic,
- deterministic logging.
curl-http
Use for protocol-level evidence:
- baseline and post-switch HTTP checks,
403/geo-block signature capture,
- header and status comparisons.
wireguard
Use when deterministic profile-based tunnels are required:
- controlled profile activation,
- route and AllowedIPs sanity expectations,
- DNS handling awareness in tunnel config.
tailscale
Use for tailnet and exit-node path:
tailscale up --exit-node=<node>,
- connectivity validation via
tailscale ping/status,
- fast fallback among available exit nodes.
dns
Use for DNS leak and propagation sanity guidance:
- resolver checks,
- authoritative vs cached record reasoning,
- explicit leak-risk interpretation when DNS path remains local.
ipinfo
Use for geo-attestation:
- validate post-switch country/region/ASN,
- compare with baseline,
- provide confidence level for geo-alignment.
moltguard
Use as prompt/tool security guardrail:
- sanitize sensitive prompt/tool content,
- detect prompt-injection patterns in fetched content,
- reduce accidental secret leakage in workflow logs.
Important limitation:
- MoltGuard is not a VPN manager and not a full network leak detector.
Canonical Causal Signal Chain
Block Detection
- baseline request to blocked endpoint,
- classify as
geo_block, ip_block, auth_block, or other_http_error.
Baseline Snapshot
- capture pre-switch public IP, country, and resolver context.
Tunnel Path Selection
- choose one path:
- provider CLI,
- WireGuard profile,
- Tailscale exit node.
- verify binary/auth/profile availability before connect.
Tunnel Activation
- connect selected path,
- confirm session state from tool output,
- enforce kill-switch preference if available.
Geo and IP Verification
- compare pre/post public IP,
- verify target country best-effort (
ipinfo.io + optional token),
- record confidence if country mismatches.
DNS Safety Check
- check resolver behavior and detect obvious DNS bypass patterns,
- flag risk if DNS appears untunneled in full-tunnel expectation.
Access Retest
- retry blocked endpoint,
- compare HTTP status/content signatures against baseline.
Task Resumption
- if retest passes, resume blocked workflow automatically (
switch-and-resume mode),
- otherwise rotate endpoint/profile once within retry budget and stop with evidence.
Suggested verification commands:
curl -s ifconfig.me
curl -s https://ipinfo.io/json
curl -I "${BLOCKED_URL}"
Leak and Safety Checks
Minimum checks before success:
- public IP changed,
- target country aligned (or deviation explicitly explained),
- endpoint moved from blocked to reachable/expected-auth state,
- DNS path does not contradict tunnel expectations,
- no unresolved high-risk MoltGuard warning (if enabled).
If kill-switch is required but not supported/verified:
- return
Needs Review and avoid high-risk task resumption.
Output Contract
Always return:
BlockDiagnosis
- block type
- baseline HTTP evidence
TunnelPath
- selected path and rationale
- provider/profile/exit node
TunnelStatus
- connect state
- pre/post IP
- target region match
DNSSafety
- resolver observation
- leak risk assessment (
low|medium|high)
SecurityStatus
- MoltGuard mode (
enabled, gateway-only, disabled)
- unresolved warnings
AccessRetest
- post-switch result
- improvement vs baseline
TaskResumption
- resumed or blocked
- reason
NextActions
- exact commands or account steps for unresolved blockers
Quality Gates
Before final output, verify:
- diagnosis is evidence-based,
- pre/post network evidence is present,
- retry count respected,
- missing credentials/keys clearly disclosed,
- provider/path limitations explicitly stated.
If any gate fails, return Needs Revision with concrete missing checks.
Failure Handling
- Missing tunnel binary/profile: return
MissingCredentials with concrete install/profile steps.
- Missing VPN account/auth session: return
MissingCredentials, skip switching stage.
- Missing
MOLTGUARD_API_KEY in detection mode: return MissingAPIKeys, continue with gateway-only or disabled mode.
- Tunnel connected but geo mismatch persists: one bounded retry with different endpoint/profile, then stop.
- Endpoint still blocked after retry: return full evidence bundle and manual-decision path.
Guardrails
- Never claim legal or terms compliance on behalf of user.
- Never claim secure state without pre/post verification.
- Never unbounded-loop region hopping.
- Never hide ambiguous or failed access states.
1---2name: vpn3description: Meta-skill for secure network tunnel setup, geo-access diagnostics, and leak-aware task resumption by orchestrating shell-scripting, curl-http, wireguard, tailscale, dns, ipinfo, and moltguard. Use when users need controlled VPN switching, region verification, DNS safety checks, and automatic retry of previously blocked workflows.4homepage: https://clawhub.ai5user-invocable: true6disable-model-invocation: false7metadata:8 source: https://skillmd.com/skills/johnalbertini14-glitch/vpn9 openclaw:10 emoji: 🔐11 requires:12 bins:13 - bash14 - curl15 anyBins:16 - nordvpn17 - mullvad18 - expressvpn19 - wg20 - tailscale21 env: []22 config: []23 note: "Requires at least one tunnel path (provider CLI, WireGuard, or Tailscale exit node). Optional security/geo enrichment: MoltGuard and IPinfo."24---2526# Purpose2728Establish a secure, verified path when access is blocked by geo/IP policy, then resume the blocked workflow safely and audibly.2930Primary outcomes:311. detect and classify block behavior,322. switch to a valid tunnel path with explicit user consent,333. verify public IP, region, and DNS safety posture,344. re-run blocked task with bounded retries,355. return an auditable connection report.3637This is an orchestration skill. It does not guarantee legal access to restricted services.3839# Required Installed Skills4041Core diagnostics/orchestration:42- `shell-scripting` (inspected latest: `1.0.0`)43- `curl-http` (inspected latest: `1.0.0`)4445Tunnel path options (at least one):46- provider CLI path (NordVPN / Mullvad / ExpressVPN) via shell orchestration47- `wireguard` (inspected latest: `1.0.0`)48- `tailscale` (inspected latest: `1.0.0`)4950Safety and verification extensions:51- `dns` (inspected latest: `1.0.0`)52- `ipinfo` (inspected latest: `1.0.0`)53- `moltguard` (inspected latest: `6.0.2`, optional but recommended)5455Install/update:5657```bash58npx -y clawhub@latest install shell-scripting59npx -y clawhub@latest install curl-http60npx -y clawhub@latest install wireguard61npx -y clawhub@latest install tailscale62npx -y clawhub@latest install dns63npx -y clawhub@latest install ipinfo64npx -y clawhub@latest install moltguard65npx -y clawhub@latest update --all66```6768Verify:6970```bash71npx -y clawhub@latest list72```7374# Required Credentials and Access7576Required access:77- valid account/session for selected tunnel path78- local executable for selected path (`nordvpn`/`mullvad`/`expressvpn` or `wg` or `tailscale`)7980Optional keys:81- `MOLTGUARD_API_KEY` (if MoltGuard remote detection mode is enabled)82- `IPINFO_TOKEN` (optional, higher quota geolocation verification)8384Preflight:8586```bash87command -v nordvpn || command -v mullvad || command -v expressvpn || command -v wg || command -v tailscale88echo "$MOLTGUARD_API_KEY" | wc -c89echo "$IPINFO_TOKEN" | wc -c90```9192Mandatory behavior:93- Never fail silently on missing keys/auth.94- Always return `MissingAPIKeys` and/or `MissingCredentials` with blocked stages.95- Continue with non-blocked diagnostics and mark output as `Partial` when needed.9697# Compliance Gate (Mandatory)9899Before any tunnel switch, confirm and record:100- user authorization to modify network routing,101- acknowledgment of legal/terms responsibility,102- stated purpose for geo-switch (testing, parity checks, privacy hardening).103104If acknowledgment is missing:105- do not execute switching commands,106- return diagnostics-only output.107108# Inputs the LM Must Collect First109110- `blocked_url` or `blocked_endpoint`111- `blocked_task_name` (example: `prediction-market-arbitrage`)112- `target_region`113- `tunnel_path` (`provider-cli`, `wireguard`, `tailscale-exit-node`)114- `provider_or_profile` (provider name, WG profile, or exit-node name)115- `risk_mode` (`diagnose-only`, `switch-and-verify`, `switch-and-resume`)116- `kill_switch_required` (`yes/no`)117- `max_retries` (default: 2)118119Do not execute switching before tunnel path and target region are explicit.120121# Tool Responsibilities122123## shell-scripting124125Use as control plane:126- executable detection,127- connect/disconnect wrappers,128- retry and cleanup logic,129- deterministic logging.130131## curl-http132133Use for protocol-level evidence:134- baseline and post-switch HTTP checks,135- `403`/geo-block signature capture,136- header and status comparisons.137138## wireguard139140Use when deterministic profile-based tunnels are required:141- controlled profile activation,142- route and AllowedIPs sanity expectations,143- DNS handling awareness in tunnel config.144145## tailscale146147Use for tailnet and exit-node path:148- `tailscale up --exit-node=<node>`,149- connectivity validation via `tailscale ping`/status,150- fast fallback among available exit nodes.151152## dns153154Use for DNS leak and propagation sanity guidance:155- resolver checks,156- authoritative vs cached record reasoning,157- explicit leak-risk interpretation when DNS path remains local.158159## ipinfo160161Use for geo-attestation:162- validate post-switch country/region/ASN,163- compare with baseline,164- provide confidence level for geo-alignment.165166## moltguard167168Use as prompt/tool security guardrail:169- sanitize sensitive prompt/tool content,170- detect prompt-injection patterns in fetched content,171- reduce accidental secret leakage in workflow logs.172173Important limitation:174- MoltGuard is not a VPN manager and not a full network leak detector.175176# Canonical Causal Signal Chain1771781. `Block Detection`179- baseline request to blocked endpoint,180- classify as `geo_block`, `ip_block`, `auth_block`, or `other_http_error`.1811822. `Baseline Snapshot`183- capture pre-switch public IP, country, and resolver context.1841853. `Tunnel Path Selection`186- choose one path:187 - provider CLI,188 - WireGuard profile,189 - Tailscale exit node.190- verify binary/auth/profile availability before connect.1911924. `Tunnel Activation`193- connect selected path,194- confirm session state from tool output,195- enforce kill-switch preference if available.1961975. `Geo and IP Verification`198- compare pre/post public IP,199- verify target country best-effort (`ipinfo.io` + optional token),200- record confidence if country mismatches.2012026. `DNS Safety Check`203- check resolver behavior and detect obvious DNS bypass patterns,204- flag risk if DNS appears untunneled in full-tunnel expectation.2052067. `Access Retest`207- retry blocked endpoint,208- compare HTTP status/content signatures against baseline.2092108. `Task Resumption`211- if retest passes, resume blocked workflow automatically (`switch-and-resume` mode),212- otherwise rotate endpoint/profile once within retry budget and stop with evidence.213214Suggested verification commands:215216```bash217curl -s ifconfig.me218curl -s https://ipinfo.io/json219curl -I "${BLOCKED_URL}"220```221222# Leak and Safety Checks223224Minimum checks before success:225- public IP changed,226- target country aligned (or deviation explicitly explained),227- endpoint moved from blocked to reachable/expected-auth state,228- DNS path does not contradict tunnel expectations,229- no unresolved high-risk MoltGuard warning (if enabled).230231If kill-switch is required but not supported/verified:232- return `Needs Review` and avoid high-risk task resumption.233234# Output Contract235236Always return:237238- `BlockDiagnosis`239 - block type240 - baseline HTTP evidence241242- `TunnelPath`243 - selected path and rationale244 - provider/profile/exit node245246- `TunnelStatus`247 - connect state248 - pre/post IP249 - target region match250251- `DNSSafety`252 - resolver observation253 - leak risk assessment (`low|medium|high`)254255- `SecurityStatus`256 - MoltGuard mode (`enabled`, `gateway-only`, `disabled`)257 - unresolved warnings258259- `AccessRetest`260 - post-switch result261 - improvement vs baseline262263- `TaskResumption`264 - resumed or blocked265 - reason266267- `NextActions`268 - exact commands or account steps for unresolved blockers269270# Quality Gates271272Before final output, verify:273- diagnosis is evidence-based,274- pre/post network evidence is present,275- retry count respected,276- missing credentials/keys clearly disclosed,277- provider/path limitations explicitly stated.278279If any gate fails, return `Needs Revision` with concrete missing checks.280281# Failure Handling282283- Missing tunnel binary/profile: return `MissingCredentials` with concrete install/profile steps.284- Missing VPN account/auth session: return `MissingCredentials`, skip switching stage.285- Missing `MOLTGUARD_API_KEY` in detection mode: return `MissingAPIKeys`, continue with gateway-only or disabled mode.286- Tunnel connected but geo mismatch persists: one bounded retry with different endpoint/profile, then stop.287- Endpoint still blocked after retry: return full evidence bundle and manual-decision path.288289# Guardrails290291- Never claim legal or terms compliance on behalf of user.292- Never claim secure state without pre/post verification.293- Never unbounded-loop region hopping.294- Never hide ambiguous or failed access states.