Google Workspace CLI (gws)
googleworkspace/cliA command-line tool for Drive, Gmail, Calendar, Sheets, Docs and Chat that builds its commands from Google's Discovery Service, plus about 40 thin agent skills.
Reads, creates and edits Lark/Feishu cloud documents and wiki pages, handles images and attachments, and works with mind notes and whiteboards through lark-cli.
A well-scoped document skill from the official Lark CLI team, with careful routing and references; it needs lark-cli set up and acts under your user identity.
lark-doc is the document skill in the official Lark CLI pack. It routes a request by scenario and then points the agent to one reference file at a time, with an explicit rule not to read all the references at the start. It reads and summarizes documents (docs +fetch), creates new ones either from scratch through a full creation workflow or by importing complete content (+create), and edits them with +update for rewrites, polishing, restructuring and layout. Helper commands parse a document URL or token and local XML and report word counts and character problems (+script), list and revert version history, insert local images or files at the end of a document, preview or download embedded media, and manage the Docx cover. Whiteboards and mind notes have their own workflows; updating an existing whiteboard must reuse its token and never create a blank one.
The skill requires the lark-cli binary and the shared lark-shared skill for authentication, identity and scope handling, which it reads only after an auth or scope error. It recommends passing --as user explicitly and gives a rule for local file paths. Comments, Drive-level copy and export, and spreadsheet data are left to other skills. The SKILL.md is only about 4 KB and written in Chinese, backed by 15 reference files of roughly 80 KB in total.
A 4 KB router SKILL.md plus 15 reference files totaling about 80 KB covering fetch, create, update, history, media, whiteboard and mind notes, though no scripts or tests ship in the skill folder.
lark-cli config risk-control off. The skill itself adds no telemetry.A Markdown-only skill folder (SKILL.md plus 15 reference files) telling the agent how to use lark-cli to work with Lark/Feishu docs and wiki pages as you. It has no scripts or allowed-tools pre-approvals; the CLI that does the work, its authorization and its default-on risk-control signals (OS type and hardware model) are outside this skill folder.
lark-cli docs and lark-cli mindnotes commands under your Lark/Feishu identity.SKILL.md:4The skill folder (16 files) on Oct 29, 2026. Not a full review: it relies on code outside the repository: lark-cli binary (@larksuite/cli) and the lark-shared skill it depends on for auth and scopes.
NVIDIA SkillSpector static scan: 1 pattern hits (analysis-evasion 1). The single SkillSpector hit is a scanner note that one referenced file (references/lark-doc-media-insert.md) was only partly analysed; the file is a Markdown guide in the skill folder that we have.
A review, not a certification. Skills change after we check them — read the files before installing.
MIT — commercial use allowed.
Not really. It is an interactive document-editing guide that expects the agent to follow a creation workflow and confirm edits with the person.
A command-line tool for Drive, Gmail, Calendar, Sheets, Docs and Chat that builds its commands from Google's Discovery Service, plus about 40 thin agent skills.
Six skills by Obsidian's Steph Ango (kepano) for Obsidian Markdown, Bases, JSON Canvas, the Obsidian CLI, Defuddle and Knap.
Anthropic's reference skills repo: 20 skills including document tools (docx, pptx, xlsx, pdf), frontend-design, skill-creator and mcp-builder.
lark-doc is a skill in the official larksuite/cli repo that lets an agent read, create and edit Lark/Feishu documents and wiki pages, including images, history and whiteboards, by running lark-cli commands.
Install the CLI with npx @larksuite/cli@latest install, configure it with lark-cli config init and log in with lark-cli auth login --recommend. Add the skills with npx skills add larksuite/cli -y -g, as the README describes.
The skill text and many references are in Chinese, but the commands and document content can be in any language the model handles; check that your model follows the Chinese routing rules reliably.
lark-doc targets Lark and Feishu documents through lark-cli, with its own XML document format and whiteboard support. For Google Docs you would use the gws-docs skills in googleworkspace/cli instead.