agentleFS
laojin1900laojin1900/365Skill/skills/gws-workspace/SKILL.md

gws-workspace

Skill1 starsChanged 4 months agoApache-2.0View on GitHub

Operate Google Workspace — Gmail, Calendar, Drive, Sheets, Docs, Slides, Chat, Tasks — through the gws CLI (github.com/googleworkspace/cli) with read-by-default safety gates. Use when the user mentions gws, Gmail, email inbox/send/reply/triage, 邮件/收件箱/发邮件/未读邮件, Google Calendar, 日程/日历/创建会议, Google Drive, 谷歌网盘/上传 Drive, Google Sheets, 谷歌表格, Google Docs, 谷歌文档, or asks for Google Workspace automation such as daily/weekly mail+calendar digests (邮件汇总/日程周报). Also use when installing gws, configuring its OAuth client, or troubleshooting gws authentication errors.

  • Installs packages

1 other project uses a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

2 rules. About 0 are common, 1 is distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Before the first gws call in a session:

    line 16
  2. Very specific:

    Confirm externally visible actions: Before gmail +send/+reply/+forward, chat +send, or creating calendar events that invite guests, restate …

    line 74

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

gws Workspace Operations

Operate a user's Google Workspace account through the gws CLI — one binary covering every Workspace API, built dynamically from Google's Discovery Service, with structured JSON output. It is open source (Apache-2.0) but not an officially supported Google product and is pre-v1: expect breaking changes and pin versions in production.

Preflight

Before the first gws call in a session:

gws auth status --format json
  • auth_method: "oauth2" or service_account → proceed.
  • auth_method: "none" → stop. Follow references/setup-guide.md with the user to install gws and complete OAuth. Never invent client IDs, client secrets, or tokens; OAuth setup requires real user actions in the Google Cloud Console.
  • command not found → install: npm install -g @googleworkspace/cli (alternatives: brew install googleworkspace-cli, cargo, nix).

The stderr line Using keyring backend: keyring is informational, not an error. Parse stdout only.

Command Model

Two surfaces; prefer helper commands when one fits.

Helper commands (+ prefix, hand-written):

ServiceCommandPurpose
gmail+send --to A --subject S --body BSend email
gmail+reply --message-id ID --body BReply (threads automatically)
gmail+triageUnread inbox summary
gmail+watchStream new mail as NDJSON
calendar+agenda [--today] [--timezone TZ]Upcoming events
calendar+insertCreate an event
drive+upload ./file --name NUpload a file
sheets+read --spreadsheet IDRead values
sheets+append --spreadsheet ID --values "a,b"Append a row
docs+writeAppend text to a document
chat+sendMessage a space
workflow+standup-report, +weekly-digest, +meeting-prepComposite digests

Discovery commands (full API surface):

gws <service> <resource> <method> --params '{"k":"v"}'   # query/path params
gws <service> <resource> <method> --json '{"k":"v"}'     # request body
gws schema gmail.users.messages.list                     # inspect any method
gws gmail --help                                         # live command list
  • Message metadata: request format: "metadata" plus metadataHeaders: ["Subject","From","Date"]; body parts are base64url in payload.body.data / payload.parts.
  • Gmail search uses standard q operators (is:unread, newer_than:7d, after:2026/07/01 before:2026/07/08).
  • Useful flags: --format json|table|yaml|csv, --page-all (NDJSON pagination), --output <path> (binary download), --upload <path>.

Safety Gates

  • Read freely. list/get/triage/agenda/read need no confirmation.
  • Confirm externally visible actions. Before gmail +send/+reply/+forward, chat +send, or creating calendar events that invite guests, restate the recipients, subject, and body summary; execute only after explicit approval.
  • Confirm destructive actions. Deleting mail, Drive files, or events, and modifying existing events or documents, require explicit approval with the target ID shown.
  • Least privilege. Request only the scopes the task needs during gws auth login; start read-only and expand. Never print credentials, tokens, or ~/.config/gws/ contents into chat output.
  • Timezone: +agenda uses the Google account timezone; pass --timezone explicitly when the user expects another zone.

Troubleshooting

SymptomCauseFix
Access blocked / access_deniedConsent screen in testing mode, account not a test userAdd the account under Google Auth Platform → Audience → Test users
Token refresh failsStale encrypted credentialsDelete ~/.config/gws/credentials.enc, re-run gws auth login
Scope/permission errorsScope not granted at loginRe-run gws auth login with the needed scopes (incremental)
Method not foundDiscovery document cached or new APIgws <service> --help; commands are generated at runtime from the live Discovery Service
No OAuth client / lost secretGoogle shows a client secret only once, at creationCreate a new Desktop client and download the JSON immediately — see references/setup-guide.md

References

  • references/setup-guide.md — from-zero install and OAuth setup: GCP project, consent screen, Desktop client, the secret-visible-once policy, test users, headless/CI credentials, and service accounts. Read it whenever gws auth status is not oauth2 or the user asks to install gws.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.