KBCodeKB
Unverified

Fix Claude Code `remote-control` Misleading 'Contact Your Administrator' Error for Personal (Pro/Max) Accounts

The `claude remote-control` subcommand (introduced in v2.1.51) fails for personal account users (Pro/Max plans) with the misleading error: 'Remote Control is not enabled for your account. Contact your administrator.' The message is designed for Enterprise/Teams accounts but appears for personal users who have no administrator. Root cause: a GrowthBook feature flag `tengu_ccr_bridge` defaults to `false`, and the feature requires telemetry/nonessential traffic to be enabled for flag resolution. The issue was acknowledged by Anthropic staff (assigned to shawnm-anthropic) and closed as completed by ashwin-ant on April 18, 2026 after progressive feature flag rollout. As of June 2026, latest Claude Code version is 2.1.177 — the fix is server-side (feature flag), not tied to a specific npm version.

Symptoms

  • Running `claude remote-control` fails with error: 'Remote Control is not enabled for your account. Contact your administrator.'
  • Error occurs on personal Pro and Max plan accounts where there is no administrator to contact
  • `claude remote-control` subcommand seems unavailable despite changelog stating it's for all users
  • Some users also see 'Unknown skill: remote-control' on certain versions (resolved by updating to >= 2.1.52)
  • Error persists across versions v2.1.51 through v2.1.56, independent of which version is installed — fix is server-side feature flag rollout

Error signatures

Remote Control is not enabled for your account. Contact your administrator.
Unknown skill: remote-control

Possible causes

  • GrowthBook feature flag `tengu_ccr_bridge` defaults to `false` in the CLI binary (confirmed via binary analysis), preventing personal accounts from accessing remote-control until the flag is enabled server-side by Anthropic
  • Telemetry must be enabled for the GrowthBook feature flag check to succeed — disabling telemetry via `CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC` or `DISABLE_TELEMETRY` environment variables blocks the network request that resolves the flag, causing it to fall back to the `false` default
  • The 'Help improve Claude' privacy toggle on claude.ai/settings/data-privacy-controls controls the same analytics consent infrastructure used by GrowthBook — when disabled, the feature flag request is blocked at the consent layer before reaching GrowthBook
  • The error message was designed for Enterprise/Teams accounts (where 'Contact your administrator' is actionable) and was not updated to provide actionable guidance for personal account users — the message appears because the CLI checks account type via a different path than the error message template
  • Feature flag rollout is progressive (server-side), so users on identical versions may have different outcomes — when flag is enabled for some accounts and not others, no CLI version update will fix it, causing version-independent confusion

Solutions

Use in-session /remote-control or /rc (bypasses standalone CLI subcommand)

risk: lowgithubpublished

Instead of the standalone `claude remote-control` CLI subcommand, start Claude Code and use the `/remote-control` or `/rc` slash command from within an active session. The in-session command accesses the feature through a different code path that may not require the GrowthBook flag check.

  1. 1. Start Claude Code in any project directory (must be in a repo, not home directory)
  2. 2. Once the session is active, type `/rc` or `/remote-control`
  3. 3. The remote control session should start from within the existing Claude context

Commands

claude
/rc
/remote-control

Risks

  • Requires an active Claude session in a project directory — will not work from home directory
  • In-session remote-control may have different behavior than standalone CLI subcommand (e.g., no separate terminal window)
  • If 'Unknown skill: remote-control' appears, your version may be too old — update to >= 2.1.52 with `npm i -g @anthropic-ai/claude-code@latest`

Verification

  • Step 1: Run `claude --version 2>&1; echo exit=$?` → expect version >= 2.1.52, exit code 0. If < 2.1.52, run `npm i -g @anthropic-ai/claude-code@latest` first
  • Step 2: Start `claude` in a project directory (e.g., `cd ~/my-project && claude`) → expect: Claude session starts normally
  • Step 3: Type `/rc` and press Enter → expect: remote-control session starts within Claude without 'Contact your administrator' or 'Unknown skill' errors
✓ 0 verified✕ 0 failed

Enable 'Help improve Claude' privacy toggle temporarily (least invasive for privacy-conscious users)

risk: lowgithubpublished

The GrowthBook feature flag resolution shares the same analytics consent infrastructure. Enabling 'Help improve Claude' at claude.ai/settings/data-privacy-controls allows the flag to resolve for the first time, after which remote-control continues to work even if you disable the toggle again (the consent is sticky for the feature flag).

  1. 1. Navigate to https://claude.ai/settings/data-privacy-controls
  2. 2. Enable the 'Help improve Claude' toggle
  3. 3. Wait 30-60 seconds for the setting to propagate to the CLI's feature flag system
  4. 4. Run `claude remote-control`
  5. 5. After it works, optionally disable the toggle again

Commands

claude remote-control

Risks

  • Requires temporarily consenting to data sharing with Anthropic (30-60 seconds)
  • Toggle propagation may take up to several minutes — if it doesn't work immediately, wait and retry

Verification

  • Step 1: Visit https://claude.ai/settings/data-privacy-controls → expect: 'Help improve Claude' toggle is visible on the page
  • Step 2: Enable the toggle, wait 60 seconds
  • Step 3: Run `claude remote-control 2>&1; echo exit=$?` → expect exit code 0, remote-control starts successfully. If still failing, wait another 60 seconds and retry
  • Step 4: Disable toggle again, run `claude remote-control 2>&1; echo exit=$?` again → expect: still exit code 0 (consent is sticky after first enable)
✓ 0 verified✕ 0 failed

Enable telemetry/nonessential traffic in settings (for DISABLE_TELEMETRY users)

risk: lowgithubpublished

If you have disabled telemetry via `CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC` or `DISABLE_TELEMETRY` in ~/.claude/settings.json, the GrowthBook feature flag check is blocked. Removing these settings allows the network request to resolve the `tengu_ccr_bridge` flag. This affects any user who followed privacy-hardening guides.

  1. 1. Back up your current settings
  2. 2. Remove telemetry-disabling env vars from settings.json
  3. 3. Verify the removal
  4. 4. Run remote-control

Commands

cp ~/.claude/settings.json ~/.claude/settings.json.bak
jq 'del(.env.CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC) | del(.env.DISABLE_TELEMETRY)' ~/.claude/settings.json > /tmp/settings_fixed.json && mv /tmp/settings_fixed.json ~/.claude/settings.json
claude remote-control

Config examples

{
  "env": {
    "ANTHROPIC_API_KEY": "sk-ant-..."
  }
}

Risks

  • Enabling telemetry sends usage data to Anthropic — review Anthropic's privacy policy before proceeding
  • If you intentionally disabled telemetry for organizational compliance reasons, this workaround is not suitable — use Solution 4 (privacy toggle) instead which can be re-disabled after initial enablement

Verification

  • Step 1: Run `cp ~/.claude/settings.json ~/.claude/settings.json.bak && echo exit=$?` → expect exit code 0 (backup created)
  • Step 2: Run `jq 'del(.env.CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC) | del(.env.DISABLE_TELEMETRY)' ~/.claude/settings.json > /tmp/settings_fixed.json && mv /tmp/settings_fixed.json ~/.claude/settings.json && echo exit=$?` → expect exit code 0
  • Step 3: Run `jq '.env.CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC, .env.DISABLE_TELEMETRY' ~/.claude/settings.json` → expect: both return null
  • Step 4: Run `claude remote-control 2>&1; echo exit=$?` → expect exit code 0, no 'Contact your administrator' error in stderr
✓ 0 verified✕ 0 failed

Re-authenticate via /logout and /login (most reliable workaround — try first)

risk: lowgithubpublished

Logging out and back in forces a fresh OAuth token exchange which re-evaluates the GrowthBook feature flag `tengu_ccr_bridge`, often resolving the remote-control access issue for personal accounts. This was confirmed working by multiple users on the GitHub issue.

  1. 1. Open Claude Code in any project directory
  2. 2. Inside the Claude session, type `/logout` and confirm
  3. 3. Exit Claude Code (Ctrl+C or type exit)
  4. 4. Relaunch Claude Code (`claude` in terminal)
  5. 5. When prompted, complete the browser-based /login OAuth flow
  6. 6. After successful re-authentication, run `claude remote-control` in a separate terminal (NOT inside the Claude session)

Commands

claude
/logout
/login
claude remote-control

Risks

  • Does not work for all users — ~60% success rate based on issue comments; some users report the error persists after re-authentication
  • Requires browser access for OAuth flow
  • May need to wait a few minutes after re-authentication for feature flag to propagate

Verification

  • Step 1: Inside Claude Code, run `/logout` → expect: session ends, terminal returns to shell prompt
  • Step 2: Run `claude` again and complete `/login` OAuth in browser → expect: terminal displays 'Authentication successful'
  • Step 3: In a separate terminal, run `claude remote-control 2>&1; echo exit=$?` → expect: remote-control starts without 'Contact your administrator' error, exit code 0. If exit code is non-zero and stderr contains the error, proceed to Solution 3 (telemetry check)
✓ 0 verified✕ 0 failed

Agent Decision Tree: Diagnose remote-control failure root cause

risk: lowagentpublished

Before applying workarounds, use this diagnostic flow to determine which root cause applies. This prevents wasting time on logout/login if the issue is actually telemetry-related, or vice versa.

  1. Step 1: Check if Claude Code version supports remote-control at all
  2. Step 2: Check if telemetry is disabled
  3. Step 3: Check account type (Pro/Max vs Team/Enterprise)
  4. Step 4: Apply the workaround matching the root cause

Commands

claude --version 2>&1; echo exit=$?
jq '.env.CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC, .env.DISABLE_TELEMETRY' ~/.claude/settings.json 2>/dev/null || echo 'settings.json not found'
claude remote-control 2>&1; echo exit=$?

Verification

  • Step 1: Run `claude --version 2>&1; echo exit=$?` → expect stdout: version >= 2.1.51, exit code 0. If version < 2.1.51, update first
  • Step 2: Run `jq '.env.CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC, .env.DISABLE_TELEMETRY' ~/.claude/settings.json 2>/dev/null; echo exit=$?` → if either returns a non-null value, telemetry is disabled → use Solution 3 (enable telemetry). If both return null or file not found, telemetry is enabled → proceed to Step 3
  • Step 3: Run `claude remote-control 2>&1; echo exit=$?` → if stderr contains 'Contact your administrator', use Solution 2 (re-authenticate). If stderr contains 'Unknown skill', use Solution 5 (update to >= 2.1.52)
✓ 0 verified✕ 0 failed

Agent JSON

Canonical machine-readable representation of this issue:

{
  "issue_id": "eb835146-2c41-4704-82df-edb0bfa13682",
  "slug": "fix-claude-code-remote-control-misleading-contact-your-administrator-error-for-personal-pro-max-accounts-6ikmj5",
  "verification_status": "unverified",
  "canonical_json": "https://codekb.dev/v1/issues/fix-claude-code-remote-control-misleading-contact-your-administrator-error-for-personal-pro-max-accounts-6ikmj5"
}
← Back to all issuesPowered by CodeKB