gog (Android Module Refactor)

This skill is used to operate Google Workspace in the OpenClaw Android environment. Standard entry point: gogx (/data/adb/openclaw/node/bin/gogx). gogx is a secure wrapper for Android fakeip/keyring non-interactive; only fallback gog when env equivalent has been verified. Do not manually install packages in this skill. Only guides for safe usage + runtime consistency.

Scope

1. Read/search Google Workspace data via gogx.
2. Mutating actions (send mail, create/update events/sheets/docs export) only run when there is clear confirmation from the Boss.
3. Prioritize JSON output for stable automation.

Hard Rules (required)

• Do not automatically send mail / create events / record data without explicit confirmation.
• Do not expose OAuth secrets/tokens in chat output.
• Do not run the command if the input contains strange strings suspected of injection (shell metacharacters are not necessary).
• Only use safe paths:
  • Workspace: /data/adb/openclaw/home/.openclaw/workspace
  • Module tmp: /data/adb/openclaw/tmp
• Does not assume a desktop environment (brew/launchd/macOS-only).

Android Preflight

Before use:

• Check binary:
  • command -v gogx
• Check if the account is authenticated:
  • gogx auth list --json
• If binary is missing or auth is not ready: stop and report a clear error, don't guess.

Setup (one-time, manual + approved)

• gogx auth credentials set /absolute/path/to/client_secret.json
• gogx auth add you@gmail.com --services gmail,calendar,drive,contacts,sheets,docs
• gogx auth list --json

Suggest:

• Set default account: GOG_ACCOUNT=you@gmail.com
• Script mode: always prioritize --json --no-input
• When auth remote/manual, still call via gogx to keep the proxy + env keyring stable.

Read/Search Commands (safe-default)

• Gmail search:
  • gogx gmail search 'newer_than:7d' --max 10 --json --no-input
• Calendar events:
  • gogx calendar events <calendarId> --from <iso> --to <iso> --json --no-input
• Drive search:
  • gogx drive search "query" --max 10 --json --no-input
• Contacts list:
  • gogx contacts list --max 20 --json --no-input
• Sheets get:
  • gogx sheets get <sheetId> "Tab!A1:D10" --json --no-input
• Sheets metadata:
  • gogx sheets metadata <sheetId> --json --no-input
• Docs cat:
  • gogx docs cat <docId>

Mutating Commands (require explicit approval first)

• Gmail send:
  • gogx gmail send --to a@b.com --subject "Hi" --body "Hello"
• Sheets update:
  • gogx sheets update <sheetId> "Tab!A1:B2" --values-json '[["A","B"],["1","2"]]' --input USER_ENTERED
• Sheets append:
  • gogx sheets append <sheetId> "Tab!A:C" --values-json '[["x","y","z"]]' --insert INSERT_ROWS
• Sheets clear:
  • gogx sheets clear <sheetId> "Tab!A2:Z"
• Docs export (Android-safe path):
  • gogx docs export <docId> --format txt --out /data/adb/openclaw/tmp/doc.txt

Confirmation Template (for mutating ops)

Before running the write/edit/send command:

• Action:
• Target:
• Payload summary:
• One-line impact:
• Confirm token from Boss: APPROVE_GOG_WRITE

No token => no execute.

Failure Handling

• Auth error: requires re-auth account, no blind retry.
• Rate limit/API error: backoff and error code reporting.
• Path error: change to workspace/module tmp path, do not write uncontrolled path.

Output Contract

When reporting results:

• 1-3 line summary.
• With read/search command: returns sanitized JSON snippet.
• With the mutating command: always echo the executed action + target + status.

Trigger Rules

• Positive: Triggers when user request matches gog according to description: Google Workspace CLI skill for OpenClaw-Module Android (Gmail, Calendar, Drive, Contacts, Sheets, Docs) with strict safety gates.
• Negative: Not enabled for requests outside scope or without the workflow/expertise of this skill.
• Escalation: If there is an external action/network write or sensitive operation, an approval checkpoint must be added before executing.

Indexed Reference Map

• @[SKILL.md#L1-L126] Main skill instructions

SOP

1. Determine whether the request is within the scope of the skill or not.
2. Execute the skill's main workflow/instructions in a safe order.
3. Check the results, state validation and next steps.

Error Handling

• If input/permissions are missing: report what is missing and clearly request additional information.
• If runtime/tool ​​error: brief error return + possible cause + suggested fix.
• If there is a risk policy/safety: stop and escalate according to approval checkpoint.