飞书Bot (`feishu-bot`)

飞书Bot is a Rust project for using a Feishu/Lark custom app as a local bot. The Cargo package is feishu-bot; the installed command is feishu-bot.

It reuses these environment variables:

FEISHU_APP_ID=cli_xxx
FEISHU_APP_SECRET=xxx
FEISHU_USER_ID=ou_xxx
# Optional, only for APIs that Feishu requires to run as a user, such as task
# list, docs/message search, minutes search, and mail send.
FEISHU_USER_ACCESS_TOKEN=u_xxx
FEISHU_REFRESH_TOKEN=r_xxx
FEISHU_OAUTH_REDIRECT_URI=http://localhost:8080/callback
# Optional, makes `feishu-bot doc create` publish new docs into Wiki by default.
FEISHU_WIKI_SPACE_ID=wiki_space_id
FEISHU_WIKI_PARENT_NODE_TOKEN=wik_xxx
FEISHU_DOC_CREATE_WIKI_DEFAULT=true
# Optional, overrides the local high-level office workflow state file.
FEISHU_OFFICE_STATE_FILE=private/office-projects.json
# With FEISHU_DOC_CREATE_WIKI_DEFAULT=true, doc create keeps the fallback docx
# and returns wiki_move_error if Wiki moving is blocked. Use --wiki-strict to fail.
# Optional, only for Helpdesk APIs.
FEISHU_HELPDESK_ID=123456
FEISHU_HELPDESK_TOKEN=ht_xxx

The CLI loads environment variables from the process, project .env, current-directory .env, and optional FEISHU_ENV_FILE or LARK_ENV_FILE. Examples that pass "$FEISHU_USER_ID" still need that variable exported in the shell, or you can pass the concrete user ID directly.

When dogfood verify reports missing_user_token, start the OAuth v2 flow. When it reports expired_user_token, try oauth refresh first and rerun the probe; if refresh fails, rerun the OAuth URL/token flow. For automated checks, dogfood verify --auto-refresh-user-token can refresh, save, and retry expired user-token probes in one run.

feishu-bot oauth url --scope offline_access --scope auth:user.id:read --scope task:task:read
feishu-bot browser open --url "<authorization_url>"
feishu-bot oauth token --code <code> --code-verifier <code_verifier> --save-env --env-file private/local.env
feishu-bot oauth refresh --save-env --env-file private/local.env
feishu-bot oauth user-info
feishu-bot dogfood verify --profile office --auto-refresh-user-token --strict --json

Code layout

The bot is split by module instead of keeping all logic in one file:

src/lib.rs owns the shared application entrypoint and JSON error handling.
src/bin/feishu-bot.rs is the primary command entrypoint.
src/bin/feishuBot.rs and src/bin/feishu.rs are compatibility aliases.
src/app/mod.rs wires the app modules and handles only the outer no-config/no-API command dispatch. src/app/dispatch.rs routes configured API commands, src/app/doctor.rs prints masked connectivity diagnostics, and src/app/response.rs centralizes Feishu JSON/binary response parsing.
src/app/cli/ contains Clap command and argument definitions, split into generated parts plus focused workflow submodules such as the office project and interaction argument groups.
src/app/office.rs contains the AI-first dispatcher plus bootstrap workflow; src/app/office/ contains report/progress workflows, interaction/status workflows, local dry-run/list helpers, document writers, resource writers, formatting, links, readback, and state helpers. src/app/office/resources/ splits bootstrap resources by chat, Wiki, Base, chat tabs, and summary/pin messages. src/app/office/local/ splits API-free list, dry-run, status, cleanup, and local dispatch helpers.
src/app/client.rs contains the Feishu HTTP client type; src/app/client/ contains token/auth/request execution plus document, IM/chat, Board, Drive, and Minutes convenience methods. src/app/client/request/ splits token/auth, common JSON wrappers, raw JSON execution, Helpdesk, binary, and multipart request helpers.
src/app/common.rs re-exports shared helpers; src/app/common/ contains query/path helpers, content and key-value input readers, JSON body helpers, and ID/content-type argument resolvers.
src/app/config.rs contains .env loading, base URL selection, and masking helpers.
src/app/approval.rs contains Approval execution; src/app/approval/ contains Approval body builders and query helper builders.
src/app/attendance.rs contains Attendance execution, group/shift/schedule/task/flow/stats operations, and attendance body builders.
src/app/base.rs contains Base/Bitable command execution; src/app/base/ contains schema, field schema, record, and permission execution, media, reference, and helper builders split by query, view, and permission domains. src/app/base/field_schema/ separates AI-friendly table field specs from OpenAPI field body construction. src/app/base/records/ separates record field input parsing, batch record input parsing, and date-field normalization.
src/app/board.rs contains Board/whiteboard execution, Mermaid/PlantUML import, raw node creation, and Board node JSON normalization.
src/app/calendar.rs contains Calendar command execution; src/app/calendar/ contains Calendar event, attendee, and free/busy body builders.
src/app/chat.rs contains Chat execution; src/app/chat/ contains chat create/update, tab/menu, and member body/query builders.
src/app/doc.rs contains docx command execution; src/app/doc/ contains Markdown/local block mapping helpers, docx media insertion helpers, raw descendant body normalization, and AI-ready raw block templates split by support matrix, child blocks, and descendant blocks.
src/app/dogfood.rs contains dogfood command dispatch plus write/message loop probes; src/app/dogfood/ contains publish execution, read-probe verify orchestration, read-probe specs, probe classification, summary aggregation, and AI remediation helpers. src/app/dogfood/probes/ splits probe output, error classification, remediation, and module-filter helpers.
src/app/drive.rs contains the Drive command dispatcher; src/app/drive/ contains Drive execution modules for file, media, comment, version, subscription, permission, and import/export operations. Query/body builders stay in focused helper modules such as src/app/drive/comment.rs, src/app/drive/permissions.rs, src/app/drive/transfer.rs, and src/app/drive/helpers/.
src/app/helpdesk.rs contains Helpdesk execution, ticket/message/FAQ operations, service-start body builders, and helpdesk message body builders.
src/app/oauth.rs contains OAuth command dispatch; src/app/oauth/ contains authorization URL/PKCE helpers, token exchange/refresh/user-info requests, token masking/env persistence, and human output formatting.
src/app/raw_api.rs contains raw Feishu API passthrough execution.
src/app/search.rs contains Search execution, doc/message search, and custom search connector body builders.
src/app/setup.rs contains setup command dispatch; src/app/setup/ contains first-run plan/scope builders, quickstart/auto/open-scopes flows, doctor/Wiki probes, Wiki bot membership automation, and browser open helpers.
src/app/sheet.rs contains Sheets execution; src/app/sheet/ contains sheet tab/style/value body builders.
src/app/task.rs contains Task command execution; src/app/task/ contains tasklist and structure execution, Task body/query helpers, input/time/reminder normalization helpers, and tasklist/comment/member collaboration builders plus section and reminder/dependency relation builders. src/app/task/helpers/ splits task create/update bodies, shared query/request helpers, member builders, input normalization, collaboration, and relations. src/app/task/custom_field/ splits custom-field metadata, option, setting, and value body builders.
src/app/vc.rs contains VC execution and meeting/reserve/recording/report/room routing; src/app/vc/ contains VC request body builders.
src/app/wiki.rs contains Wiki/knowledge-space command execution; src/app/wiki/ contains typed Wiki request body builders, and src/app/wiki/route/ splits route-check calls, write-probe publishing, recommendation text, and the shared Wiki request wrapper.
src/app/help.rs exposes AI help text; src/app/help/ splits long AI-facing help sections by workflow area. src/app/manifest.rs builds the machine-readable AI manifest; src/app/manifest/ splits module metadata by workflow area. src/app/scopes.rs prints and filters scope grants, with the static scope registry split by domain in src/app/scopes/groups/; src/app/output.rs re-exports output helpers, while src/app/output/ contains generic response summaries, document block/code output, and Feishu block/code label maps.
src/app/mail.rs contains Mail execution, mailbox auth selection, message/folder/contact/settings/rule/label operations, and mail send body builders.
src/app/message.rs contains IM message execution, image/file/video helpers, reactions, and pins; src/app/message/ contains voice synthesis/conversion send entrypoints plus voice prepare/audio/synth helpers, message content/reaction body builders, polling cursor/state, and loop-check/readback probe helpers.
src/app/minutes.rs contains Minutes execution, metadata/artifacts/media/transcript operations, token parsing, and search body builders.
src/app/okr.rs contains OKR execution, OKR query helpers, and ID validation.
src/app/people.rs contains Contact, Directory, and CoreHR execution and body builders; src/app/people/ contains Directory, CoreHR, and Hire execution and helpers. Hire-specific query/body/id-type helpers live under src/app/people/hire/.
src/app/tests.rs contains unit tests.

See docs/ARCHITECTURE.md before adding new Feishu APIs. See docs/DOGFOOD.md before claiming a capability works. See docs/QUALITY_STATUS.md before continuing broad refactors. See skills/feishu-bot-office/SKILL.md for the repeatable AI office workflow.

Quick checks

scripts/check-commit.sh
scripts/ci-local.sh
feishu-bot ai
feishu-bot manifest
feishu-bot --help
feishu-bot setup --help
feishu-bot setup plan
feishu-bot dogfood verify --profile office --auto-refresh-user-token --strict --json
feishu-bot oauth --help
feishu-bot scopes --group all
feishu-bot doc capabilities
feishu-bot doc template --kind all
feishu-bot doc template --kind support-matrix
feishu-bot board --help
feishu-bot contact --help
feishu-bot directory --help
feishu-bot chat --help
feishu-bot message --help
feishu-bot message list --help
feishu-bot message reaction --help
feishu-bot message pin --help
feishu-bot base --help
feishu-bot base media --help
feishu-bot task --help
feishu-bot drive --help
feishu-bot calendar --help
feishu-bot vc --help
feishu-bot minutes --help
feishu-bot search --help
feishu-bot okr --help
feishu-bot attendance --help
feishu-bot mail --help
feishu-bot corehr --help
feishu-bot helpdesk --help
feishu-bot hire --help
feishu-bot wiki --help
feishu-bot sheet --help
feishu-bot approval --help
feishu-bot api --help
feishu-bot doctor
feishu-bot office --help
feishu-bot office list
feishu-bot office bootstrap --project "demo" --dry-run
feishu-bot office report --project "demo" --title "dry run" --content "hello" --dry-run
feishu-bot office status --project "demo"
feishu-bot dogfood verify --profile office --auto-refresh-user-token --strict --json
feishu-bot --json manifest --module base
feishu-bot doc preview --file ./docs/feishu-bot-guide.md

The commands above are read-only or local-only except dogfood verify, which is read-only by default but may call real Feishu probes. The commands below create real Feishu data and should only be used when you intentionally want a smoke object/message:

feishu-bot office progress --project "demo" --title "status" --summary "ok"
feishu-bot dogfood publish --title "Bot smoke" --file ./demo.md

First-run setup that may operate the signed-in browser or add the app bot to a Wiki space is intentionally separate from quick checks:

feishu-bot setup quickstart --no-wiki-bot
feishu-bot setup quickstart --open-browser

Install it as a normal CLI:

cargo install --path .
feishu-bot doctor

Required app scopes for the full workflow:

im:message, im:message:readonly, im:message.history:readonly, im:message:send_as_bot, im:message:update, im:message:recall
im:message.group_at_msg:readonly, im:message.group_msg, im:message.p2p_msg:readonly
im:message.reactions:read, im:message.reactions:write_only
im:message.pins:read, im:message.pins:write_only
im:resource
im:chat, im:chat:create, im:chat:operate_as_owner, im:chat.group_info:readonly
contact:contact.base:readonly, contact:contact:access_as_app, contact:contact:readonly, contact:contact:readonly_as_app
Directory scopes from feishu-bot scopes --group directory, including directory:employee:search, directory:employee:read, directory:employee:list, and per-field employee/department/job/work-place read scopes.
docx:document, docx:document:create
docx:document.block:convert
board:whiteboard:node:create, board:whiteboard:node:read
bitable:app, bitable:app:readonly
Base scopes from feishu-bot scopes --group base, including app/table/field/view/record, dashboard, workflow, form, role/member, and media scopes
task:task:write or task:task:writeonly; task:task:read / task:task:readonly for reads
Tasklist/comment scopes from feishu-bot scopes --group task
Drive scopes from feishu-bot scopes --group drive, including drive:file, drive:file:readonly, drive:file:upload, and drive:file:download
VC scopes from feishu-bot scopes --group vc, including meeting/reserve write scopes, vc:meeting.participant:write, report, recording, meeting room, and room-level read scopes
Minutes scopes from feishu-bot scopes --group minutes, including basic read, search, AI artifacts, media export, and transcript export. minutes search also requires FEISHU_USER_ACCESS_TOKEN.
Search scopes from feishu-bot scopes --group search, including search:docs:read, search:message, search:data_source, and search:data_source:readonly. Docs/message search requires FEISHU_USER_ACCESS_TOKEN.
OKR scopes from feishu-bot scopes --group okr, including okr:okr.period:readonly, okr:okr:readonly, and okr:okr.
Attendance scopes from feishu-bot scopes --group attendance, including attendance:rule, attendance:rule:readonly, attendance:task, and attendance:task:readonly.
Mail scopes from feishu-bot scopes --group mail, including message read/send, mailbox, folder, contact, rule, and sensitive field scopes. --mailbox me and sending mail require FEISHU_USER_ACCESS_TOKEN.
CoreHR scopes from feishu-bot scopes --group corehr, including department, job, job-data, person, process, and sensitive field scopes. CoreHR also needs Feishu People data-range grants in the Open Platform.
Helpdesk scopes from feishu-bot scopes --group helpdesk, including helpdesk:all:readonly, helpdesk:all, and helpdesk:helpdesk:access. Helpdesk APIs also require FEISHU_HELPDESK_ID and FEISHU_HELPDESK_TOKEN.
Hire scopes from feishu-bot scopes --group hire, including job, talent, hire:people_cli, application, interview, process, requirement, attachment, location, subject, site, and related sensitive entity scopes. Hire also needs Feishu Hire product/data-range grants in the Open Platform.
Cloud document permission scopes from feishu-bot scopes --group permission, including docs:permission.member:* and docs:permission.setting:*
Sheet scopes from feishu-bot scopes --group sheet, including create, metadata read, and values write scopes
Calendar, Wiki, and Approval scopes prompted by their API errors

AI handoff

Future AI agents should start with:

feishu-bot ai
feishu-bot manifest
feishu-bot scopes --group all
feishu-bot scopes --group directory
feishu-bot scopes --group helpdesk
feishu-bot scopes --group hire
feishu-bot doctor

feishu-bot ai is embedded in the binary and does not need credentials. feishu-bot manifest prints a machine-readable module/scope/command manifest for future AI agents; use feishu-bot manifest --module directory or feishu-bot manifest --module helpdesk or feishu-bot manifest --module hire for compact product-specific handoffs. feishu-bot scopes prints scope groups and Open Platform grant URLs. The operator playbook covers environment, scopes, receiver IDs, smoke tests, document-writing workflow, browser bridge, and troubleshooting.

Messages

feishu-bot message send --to "$FEISHU_USER_ID" --text "hello from Feishu Bot"
feishu-bot message loop-check --to "$FEISHU_USER_ID" --to-type open-id
feishu-bot message send-json --to "$FEISHU_USER_ID" --msg-type text --content-json '{"text":"native JSON message"}'
feishu-bot message upload-image --file ./image.png
feishu-bot message send-image --to "$FEISHU_USER_ID" --file ./image.png
feishu-bot message upload-file --file ./demo.mp4 --file-type mp4 --duration 3000
feishu-bot message send-file --to "$FEISHU_USER_ID" --file ./demo.mp4 --file-type mp4
feishu-bot message send-file --to "$FEISHU_USER_ID" --file ./voice.opus --file-type opus --duration 3000
feishu-bot message list --container-id oc_xxx --container-id-type chat --page-size 20
feishu-bot message get --message-id om_xxx
feishu-bot message read-users --message-id om_xxx
feishu-bot message resource --message-id om_xxx --file-key file_xxx --type file --output ./download.bin
feishu-bot message download-image --image-key img_v2_xxx --output ./image.png
feishu-bot message download-file --file-key file_xxx --output ./download.bin
feishu-bot message reaction list --message-id om_xxx
feishu-bot message reaction add --message-id om_xxx --emoji-type SMILE
feishu-bot message reaction delete --message-id om_xxx --reaction-id <reaction_id>
feishu-bot message pin list --chat-id oc_xxx
feishu-bot message pin add --message-id om_xxx
feishu-bot message pin delete --message-id om_xxx

Use message upload-image when the AI needs an image_key; Feishu supports JPG/JPEG/PNG/WEBP/GIF/BMP/ICO/TIFF/HEIC up to 10 MB. Use message upload-file for files, audio, and videos up to 30 MB; --file-type mp4 is the native video path and stream is the fallback for other file types. send-image and send-file do the upload and message send in one step. For send-file, --msg-type auto sends mp4 as native media, opus as native audio, and all other file types as file; use --cover-image-key for video cover images when the AI already has an uploaded image key.

Use message loop-check for dogfood: it sends one message and immediately reads back the message, recent chat messages, chat metadata, chat members, and read-users. Use message send-json, reply-json, edit-json, and delete for native Feishu message payloads. Use message list/get/resource/read-users for message inspection and downloads, and message reaction/pin for message interactions.

Receiver ID type is inferred:

ou_... -> open_id
oc_... -> chat_id
on_... -> union_id
name@example.com -> email

Chats

feishu-bot chat list
feishu-bot chat search --query "项目"
feishu-bot chat get --chat-id <chat_id>
feishu-bot chat create --name "AI 项目群" --user "$FEISHU_USER_ID"
feishu-bot chat member list --chat-id <chat_id>
feishu-bot chat member add --chat-id <chat_id> --id "$FEISHU_USER_ID"
feishu-bot chat member is-in-chat --chat-id <chat_id>
feishu-bot chat member delete --chat-id <chat_id> --id <open_id>

Notifications

feishu-bot notify \
  --project feishu-bot \
  --status done \
  --task "Bot dogfood" \
  --summary "Rust CLI sent this card" \
  --details "token ok|message ok|doc API ready" \
  --text "This card was sent by the new Rust CLI."

Without --to, notify creates or reuses a project chat and stores the mapping in ~/.config/feishu/projects.json. It also reads old mappings from ~/.config/feishu/projects.json.

Documents

Create a new docx document and append content. There are two writers:

--writer local: no conversion API; maps a small Markdown-ish syntax to native docx blocks.
--writer official: calls Feishu's official Markdown/HTML -> docx block converter. Use this first for richer Markdown/HTML such as tables, inline styles, and links.

The local writer supports headings, paragraphs, unordered lists, ordered lists, code fences, quotes, todos, and dividers.

Preview the generated native blocks before writing a larger document:

feishu-bot doc preview --file ./guide.md
feishu-bot --json doc preview --file ./guide.md

feishu-bot doc create \
  --title "Feishu Bot dogfood" \
  --content $'# Hello\n\n- created from CLI\n- written through docx block API'

Use official conversion:

feishu-bot doc convert --file ./guide.md
feishu-bot dogfood publish --title "Guide dogfood" --file ./guide.md
feishu-bot doc create --writer official --title "Guide" --file ./guide.md
feishu-bot doc create --writer official --title "Guide" --file ./guide.md --send-to "$FEISHU_USER_ID" --send-loop-check
feishu-bot doc create --writer official --content-type html --title "HTML Guide" --file ./guide.html
feishu-bot doc create --writer official --title "Dogfood" --file ./demo.md --wiki --wiki-space-id <space_id> --wiki-fallback-ok
FEISHU_DOC_CREATE_WIKI_DEFAULT=true FEISHU_WIKI_SPACE_ID=<space_id> feishu-bot doc create --writer official --title "Dogfood" --file ./demo.md
FEISHU_DOC_CREATE_WIKI_DEFAULT=true FEISHU_WIKI_SPACE_ID=<space_id> feishu-bot doc create --writer official --title "Strict Wiki" --file ./demo.md --wiki-strict
feishu-bot drive import file --file ./guide.html --type docx --folder-token "" --title "HTML Guide"

HTML online preview is not exposed as “host this arbitrary HTML page” in the public Drive/docx OpenAPI. The AI-safe solution is to convert or import HTML into a native Feishu docx: use doc create --writer official --content-type html for HTML strings/files, or drive import file --type docx when the AI has a real .html file and wants Feishu's import pipeline to return a previewable online document URL.

Append to an existing document:

feishu-bot doc append --document-id doxcn... --file ./notes.md

Read plain text:

feishu-bot doc raw --document-id doxcn...

Inspect native docx block types:

feishu-bot doc blocks --document-id doxcn...

Print the AI writing boundary, including Mermaid and advanced block subtypes:

feishu-bot doc capabilities
feishu-bot doc template --kind all
feishu-bot doc template --kind support-matrix

Advanced AI raw block writing:

feishu-bot doc insert-media --document-id doxcn... --kind image --file ./image.png --width 640 --align 2
feishu-bot doc insert-media --document-id doxcn... --kind file --file ./attachment.pdf --view-type 1
feishu-bot drive media upload --parent-type docx_image --parent-node <image_block_id> --drive-route-token <document_id> --file ./image.png
feishu-bot drive media upload --parent-type docx_file --parent-node <file_block_id> --drive-route-token <document_id> --file ./attachment.pdf
feishu-bot doc template --kind image-child > image.json
feishu-bot doc template --kind file-child > file.json
feishu-bot doc template --kind iframe-child > iframe.json
feishu-bot doc template --kind link-preview-child > link.json
feishu-bot doc template --kind table-descendant > table.json
feishu-bot doc template --kind grid-descendant > grid.json
feishu-bot doc template --kind callout-descendant > callout.json
feishu-bot doc template --kind quote-container-descendant > quote-container.json
feishu-bot doc template --kind agenda-descendant > agenda.json
feishu-bot doc append-json --document-id doxcn... --file ./children.json
feishu-bot doc append-descendant --document-id doxcn... --file ./descendant-body.json

These commands are the escape hatch for rarer/newer Feishu blocks that are not modeled by the local writer yet, such as table, table_cell, grid, grid_column, iframe, image, file, sheet, bitable, callout, quote_container, add_ons, board, agenda, link_preview, wiki catalog blocks, or custom app blocks, as long as the JSON follows Feishu's docx block API and includes any required subtype fields. For normal docx image/file blocks, prefer doc insert-media: it creates the target media block, uploads the local asset with drive media upload, then patches the block with replace_image or replace_file. Use the lower-level drive media upload plus feishu-bot api route only for unusual block updates not covered by insert-media. For example, iframe uses iframe.component.type/url, tables use table.property, and callouts use callout.background_color.

Mermaid is preserved as source in a code block. Feishu's public docx OpenAPI does not expose a Mermaid source field for a rendered plugin block; diagram only exposes diagram_type for flowchart/UML and is not writable through the public docx API. For rendered Mermaid, create or locate a document board block and import the Mermaid into the whiteboard:

feishu-bot doc template --kind board-child > board.json
feishu-bot doc append-json --document-id doxcn... --file board.json
feishu-bot doc blocks --document-id doxcn...
feishu-bot board import --whiteboard-id <whiteboard_id> --syntax mermaid --file ./diagram.mmd

Mindnote is also not currently writable through public docx OpenAPI. Keep it as an existing/read block or use Feishu UI until Feishu exposes a write shape. Task blocks, source/reference synced blocks, and AI template blocks are also read-only from public docx write APIs; create tasks through feishu-bot task.

Send a document link:

feishu-bot doc send-link --document-id doxcn... --to "$FEISHU_USER_ID" --send-loop-check

Set FEISHU_DOC_BASE_URL if your tenant document URL is not https://my.feishu.cn/docx/{document_id}. It may either be a prefix or a template containing {document_id}.

Board / Mermaid

feishu-bot board import --whiteboard-id <whiteboard_id> --syntax mermaid --file ./diagram.mmd
feishu-bot board import --whiteboard-id <whiteboard_id> --syntax plantuml --file ./diagram.puml
feishu-bot board node-create --whiteboard-id <whiteboard_id> --file ./nodes.json

whiteboard_id is the board.token from a docx block with block_type=43. Use feishu-bot doc blocks --document-id <id> to inspect it after creating or locating a board block; the normal output prints board_token[block_id]=... when a board block exists.

Base / Bitable

feishu-bot base parse-url 'https://example.feishu.cn/base/<app_token>?table=<table_id>&view=<view_id>'
feishu-bot base create --name "AI 工作台"
feishu-bot base update --app-token <app_token> --name "AI 工作台 v2"
feishu-bot base copy --app-token <app_token> --name "AI 工作台副本" --folder-token <folder_token>
feishu-bot base table list --app-token <app_token>
feishu-bot base table create --app-token <app_token> --name "需求"
feishu-bot base table create --app-token <app_token> --name "需求" --default-view-name "默认视图" --field "标题:text" --field "状态:single-select:待处理:0|完成:1" --field "金额:currency:0.00|CNY" --field "截止日期:date:yyyy/MM/dd"
feishu-bot base table batch-create --app-token <app_token> --name "需求" --name "实验"
feishu-bot base table update --app-token <app_token> --table-id <table_id> --name "需求池"
feishu-bot base field list --app-token <app_token> --table-id <table_id> --view-id <view_id> --text-field-as-array
feishu-bot base field create --app-token <app_token> --table-id <table_id> --name "状态" --kind single-select --option "待处理:0" --option "完成:1"
feishu-bot base field create --app-token <app_token> --table-id <table_id> --name "金额" --kind currency --formatter "0.00" --currency-code CNY
feishu-bot base field create --app-token <app_token> --table-id <table_id> --name "截止日期" --kind date --date-formatter "yyyy/MM/dd" --auto-fill false
feishu-bot base field update --app-token <app_token> --table-id <table_id> --field-id <field_id> --name "阶段" --kind multi-select --option "进行中:2" --option "阻塞:3"
feishu-bot base view list --app-token <app_token> --table-id <table_id>
feishu-bot base view create --app-token <app_token> --table-id <table_id> --name "AI 看板" --view-type kanban
feishu-bot base view update --app-token <app_token> --table-id <table_id> --view-id <view_id> --hidden-field-id fld_internal --filter-conjunction and --filter-condition 'fld_status:3:is:json:["opt_done"]' --hierarchy-field-id fld_parent
feishu-bot base record create --app-token <app_token> --table-id <table_id> --field 标题=hello --field 分数=12.5 --field 完成=true
feishu-bot base record create --app-token <app_token> --table-id <table_id> --field '附件=json:[{"file_token":"<file_token>"}]'
feishu-bot base record create --app-token <app_token> --table-id <table_id> --fields-json '{"标题":"hello"}'
feishu-bot base record search --app-token <app_token> --table-id <table_id> --view-id <view_id> --field-name "标题" --automatic-fields
feishu-bot base record search --app-token <app_token> --table-id <table_id> --filter-json '{"conjunction":"and","conditions":[]}' --sort-json '[]'
feishu-bot base record search --app-token <app_token> --table-id <table_id> --body-json '{}'
feishu-bot base record batch-update --app-token <app_token> --table-id <table_id> --records-json '[{"record_id":"rec...","fields":{"状态":"done"}}]'
feishu-bot base record batch-create --app-token <app_token> --table-id <table_id> --record-field 0:标题=A --record-field 0:状态=open --record-field 1:标题=B
feishu-bot base record batch-update --app-token <app_token> --table-id <table_id> --record-id rec_a --record-id rec_b --record-field 0:状态=done --record-field 1:清空=null
feishu-bot base record batch-delete --app-token <app_token> --table-id <table_id> --record-id rec_xxx
feishu-bot base media upload --app-token <app_token> --kind image --file ./screenshot.png
feishu-bot base media upload --app-token <app_token> --kind file --file ./demo.mp4
feishu-bot base media field-value --file-token <file_token> --field "附件"
feishu-bot base record update --app-token <app_token> --table-id <table_id> --record-id <record_id> --field 状态=done --field 清空=null
feishu-bot base record update --app-token <app_token> --table-id <table_id> --record-id <record_id> --fields-json '{"附件":[{"file_token":"<file_token>"}]}'
feishu-bot base media tmp-url --file-token <file_token> --table-id <table_id> --field-id <field_id> --record-id <record_id>
feishu-bot base media download --file-token <file_token> --output ./asset.bin --table-id <table_id> --field-id <field_id> --record-id <record_id>
feishu-bot base dashboard list --app-token <app_token>
feishu-bot base dashboard copy --app-token <app_token> --block-id <block_id> --name "指标副本"
feishu-bot base workflow list --app-token <app_token>
feishu-bot base workflow block-list --app-token <app_token>
feishu-bot base workflow update --app-token <app_token> --workflow-id <workflow_id> --status disable
feishu-bot base form get --app-token <app_token> --table-id <table_id> --form-id <form_id>
feishu-bot base form update --app-token <app_token> --table-id <table_id> --form-id <form_id> --body-json '{...}'
feishu-bot base update --app-token <app_token> --is-advanced true
feishu-bot base role list --app-token <app_token> --api-version v2
feishu-bot base role create --app-token <app_token> --api-version v2 --name "只读成员" --table-roles-json '[...]' --allow-base-complex-edit false --allow-copy false
feishu-bot base role create --app-token <app_token> --api-version v2 --body-json '{"role_name":"只读成员","table_roles":[...],"base_rule":{"base_complex_edit":0,"copy":0}}'
feishu-bot base member list --app-token <app_token> --role-id <role_id>
feishu-bot base member add --app-token <app_token> --role-id <role_id> --member-id "$FEISHU_USER_ID" --member-id-type open_id
feishu-bot base member batch-add --app-token <app_token> --role-id <role_id> --member open_id:ou_xxx --member chat_id:oc_xxx
feishu-bot base field delete --app-token <app_token> --table-id <table_id> --field-id <field_id>
feishu-bot base table batch-delete --app-token <app_token> --table-id <table_id>
feishu-bot base table delete --app-token <app_token> --table-id <table_id>

For existing Bases, grant both Open Platform scopes and document/Base access to the app itself. Advanced permission role/member commands also require advanced permissions enabled and manageable permission on the Base. For advanced permissions 2.0 custom roles, prefer base role list/create --api-version v2. The v2 path uses /base/v2/apps/:app_token/roles and supports base_rule permission points: --allow-base-complex-edit false controls Base copy/download/print, and --allow-copy false controls content copy. Use --base-rule-json for future v2 permission points. For base record create/update --field, values are JSON-parsed when possible: 12.5, true, and null become native Base values. Use json:[...] for arrays/objects such as attachments, and str:true when the desired value is a literal string. For batch create/update, use --record-field index:name=value to group fields by record index. Batch update also needs repeated --record-id values in the same index order, unless each raw records JSON entry already has record_id. When the user pastes a Base link, run feishu-bot base parse-url first. It accepts both /base/<app_token> and /app/<app_token> Base application URLs. If it is a Wiki URL, use the returned wiki_node_token with feishu-bot wiki node; when obj_type is bitable, the returned obj_token is the Base app_token. Use base field list --view-id after parsing a Base URL so field metadata matches the user's current view. Add --text-field-as-array when an AI needs structured field descriptions instead of plain description strings. Use base view update typed flags for common view property edits: hide fields with --hidden-field-id, configure filters with --filter-conjunction and --filter-condition field_id:field_type:operator:value, and set hierarchy or sub-record structure with --hierarchy-field-id. Prefix a condition value with json: when Feishu expects the filter value to be a compact JSON string. For fields, prefer --kind instead of numeric --type. --kind covers common types: text, barcode, email, number, progress, currency, rating, single-select, multi-select, date, checkbox, user, phone, url, attachment, link, formula, duplex-link, location, group, and auto-number. --property-json remains the native escape hatch. When creating a new table, prefer repeated --field "name:kind[:config]" so the AI can create common fields in one command. Config supports select options split by |, currency formatter|CURRENCY, date formatters, formulas, linked table IDs, user/group multiple=true, location input type, and json:{...} as the raw field.property escape hatch. --fields-json remains available for full official table.fields payloads. For record reads, prefer base record search with typed flags: --view-id, repeated --field-name, --filter-json, --sort-json, and --automatic-fields. Feishu ignores view_id when filter or sort is set, because that becomes a full-table conditional query. Use --body-json only when the AI needs the exact native request body. Base image/file/video fields are two-step: upload with base media upload, then write the returned file_token into an attachment field with base record create/update. base media field-value prints the exact field JSON. For Base advanced permissions, pass table/field/record IDs to base media tmp-url or base media download so the tool builds the official bitablePerm extra.

Tasks

feishu-bot task tasklist create --name "AI 项目清单"
feishu-bot task tasklist list
feishu-bot task tasklist tasks --tasklist-guid <tasklist_guid>
feishu-bot task list --completed false --type my_tasks
feishu-bot task tasklist add-member --tasklist-guid <tasklist_guid> --editor "$FEISHU_USER_ID"
feishu-bot task tasklist remove-member --tasklist-guid <tasklist_guid> --viewer "$FEISHU_USER_ID"
feishu-bot task create --summary "follow up" --description "created by feishu-bot"
feishu-bot task create --summary "review proposal" --due-at 2026-06-02T15:00:00+08:00 --start-date 2026-06-02
feishu-bot task create --summary "submit proposal" --due-at "2026-06-03 18:00" --reminder-minute 30
feishu-bot task create --summary "all-day milestone" --due-date 2026-06-05 --mode 1 --is-milestone true
feishu-bot task create --summary "weekly sync" --due-ms 1780000000000 --due-all-day --repeat-rule "FREQ=WEEKLY;INTERVAL=1"
feishu-bot task create --summary "external ticket" --origin-json '{"platform_i18n_name":{"en_us":"AI System"},"href":{"url":"https://example.com/t/1"}}' --custom-complete-json '{"pc":{"tip":{"en_us":"Finish in source system"}}}' --extra eyJzb3VyY2UiOiJhaSJ9
feishu-bot task create --summary "milestone" --due-date 2026-06-30 --mode 1 --is-milestone true --reminder-minute 30
feishu-bot task get --guid <task_guid>
feishu-bot task update --guid <task_guid> --due-at 2026-06-03T18:00:00+08:00 --mode 1 --is-milestone true
feishu-bot task update --guid <task_guid> --clear-start --clear-repeat-rule --clear-custom-complete
feishu-bot task member add --task-guid <task_guid> --assignee "$FEISHU_USER_ID"
feishu-bot task member remove --task-guid <task_guid> --follower "$FEISHU_USER_ID"
feishu-bot task tasklists --task-guid <task_guid>
feishu-bot task add-tasklist --task-guid <task_guid> --tasklist-guid <tasklist_guid> --section-guid <section_guid>
feishu-bot task remove-tasklist --task-guid <task_guid> --tasklist-guid <tasklist_guid>
feishu-bot task section list --resource-type tasklist --resource-id <tasklist_guid>
feishu-bot task section create --resource-type tasklist --resource-id <tasklist_guid> --name "In progress"
feishu-bot task section tasks --section-guid <section_guid>
feishu-bot task custom-field list --resource-type tasklist --resource-id <tasklist_guid>
feishu-bot task custom-field create --resource-id <tasklist_guid> --name "Priority" --type single_select --option High --option Medium --option Low
feishu-bot task custom-field set-value --task-guid <task_guid> --custom-field-guid <field_guid> --type single-select --option-guid <option_guid>
feishu-bot task custom-field set-value --task-guid <task_guid> --custom-field-guid <field_guid> --type text --value "reviewed"
feishu-bot task custom-field set-value --task-guid <task_guid> --custom-field-guid <field_guid> --type member --member "$FEISHU_USER_ID"
feishu-bot task custom-field set-value --task-guid <task_guid> --custom-field-guid <field_guid> --type multi-select --option-guid <option_a> --option-guid <option_b>
feishu-bot task custom-field set-value --task-guid <task_guid> --custom-field-guid <field_guid> --type text --clear
feishu-bot task custom-field option update --custom-field-guid <field_guid> --option-guid <option_guid> --is-hidden true
feishu-bot task attachment list --resource-id <task_guid>
feishu-bot task attachment upload --resource-id <task_guid> --file ./image.png --file ./brief.pdf
feishu-bot task attachment delete --attachment-guid <attachment_guid>
feishu-bot task reminder add --task-guid <task_guid> --reminder-minute 30
feishu-bot task reminder remove --task-guid <task_guid> --reminder-id <reminder_id>
feishu-bot task dependency add --task-guid <task_guid> --dependency-task-guid <next_task_guid>
feishu-bot task dependency remove --task-guid <task_guid> --dependency-task-guid <next_task_guid>
feishu-bot task comment create --task-guid <task_guid> --content "progress note"
feishu-bot task comment list --task-guid <task_guid>
feishu-bot task comment get --comment-id <comment_id>
feishu-bot task comment update --comment-id <comment_id> --content "updated note"
feishu-bot task comment delete --comment-id <comment_id>
feishu-bot task complete --guid <task_guid>

Complex task, tasklist, section, dependency, custom-field, and comment payloads can be passed as Feishu native JSON with --body-json. Custom field values are typed through task custom-field set-value; use --clear for the official empty value. For task times, prefer --due-at/--start-at with RFC3339 values such as 2026-06-02T15:00:00+08:00, or local values such as 2026-06-02 15:00. Use --due-date/--start-date for all-day dates. --due-ms and --start-ms remain available when the AI already has Feishu millisecond timestamps. Task reminders are relative to the due time. Use --reminder-minute when creating a task/subtask or adding a reminder to an existing task. Feishu currently supports one reminder per task, so change a reminder by removing the old one and adding the new one. feishu-bot task list defaults to --auth user because Feishu's official task-list API requires user_access_token and returns the caller's "my tasks" view. Use --completed true|false to filter that view; --type defaults to my_tasks. Core task/tasklist/member/reminder/subtask commands plus section/custom-field/attachment/dependency/comment wrappers support --auth tenant|user. Use tenant auth for app-owned task data and user auth when the AI must match a Feishu user's Task Center visibility. Task attachment upload accepts one to five local files per request, each up to 50 MB.

Drive

feishu-bot drive list --folder-token <folder_token>
feishu-bot drive folder create --name "AI 输出" --folder-token ""
feishu-bot drive upload --folder-token <folder_token> --file ./report.pdf
feishu-bot drive media upload --parent-type docx_image --parent-node <image_block_id> --drive-route-token <document_id> --file ./image.png
feishu-bot drive media upload --parent-type bitable_file --parent-node <app_token> --drive-route-token <app_token> --file ./video.mp4
feishu-bot drive media tmp-url --file-token <media_token>
feishu-bot drive media download --file-token <media_token> --output ./asset.bin
feishu-bot drive import file --file ./page.html --type docx --folder-token "" --title "HTML Preview"
feishu-bot drive import get --ticket <ticket>
feishu-bot drive export file --token <docx_token> --type docx --file-extension pdf --output ./doc.pdf
feishu-bot drive export create --token <sheet_token> --type sheet --file-extension xlsx
feishu-bot drive upload-large --file ./large-video.mp4 --folder-token <folder_token>
feishu-bot drive download --file-token <file_token> --output ./report.pdf
feishu-bot drive permission public-get --token <docx_token> --file-type docx
feishu-bot drive permission public-update --token <docx_token> --file-type docx --link-share-entity tenant_readable
feishu-bot drive permission member-list --token <docx_token> --file-type docx
feishu-bot drive permission member-add --token <docx_token> --file-type docx --member-id "$FEISHU_USER_ID" --perm edit
feishu-bot drive stats --file-token <token> --file-type docx
feishu-bot drive copy --file-token <token> --file-type docx --folder-token <folder_token>

drive upload uses Feishu drive/v1/files/upload_all, so it is for Drive files. drive media upload uses drive/v1/medias/upload_all, so it is for images/files/videos attached to docx, Sheets, Base, or import staging. Both single-call upload flows are for non-empty files up to 20 MB. Use feishu-bot drive upload-large for larger Drive files; it wraps Feishu's official upload_prepare / upload_part / upload_finish multipart flow.

Use drive import file to turn local HTML/Markdown/TXT/Office files into native online Feishu docs/sheets/Base files. For HTML preview, prefer importing or converting to docx; raw uploaded .html files are just Drive files and should not be treated as a guaranteed rendered web page.

Use drive export file to create, poll, and download a cloud document export in one command. Feishu exports doc/docx to PDF or Word, and Sheets/Base to XLSX or CSV; export files must be downloaded promptly after the task finishes.

drive permission wraps the Feishu cloud document permission APIs. Use it after creating a docx/Base/sheet/file when the AI needs to share access with a user, chat, or other collaborator. Use permission member-list as the readback step after add/update/delete before claiming a recipient can access the artifact.

Calendar

feishu-bot calendar primary
feishu-bot calendar list
feishu-bot calendar freebusy list --user-id "$FEISHU_USER_ID" --time-min 2026-06-01T09:00:00+08:00 --time-max 2026-06-01T18:00:00+08:00
feishu-bot calendar freebusy batch --user-id ou_xxx --user-id ou_yyy --time-min 2026-06-01T09:00:00+08:00 --time-max 2026-06-01T18:00:00+08:00
feishu-bot calendar event create --calendar-id <calendar_id> --summary "sync" --start-ts <unix_seconds> --end-ts <unix_seconds>
feishu-bot calendar attendee add --calendar-id <calendar_id> --event-id <event_id> --user "$FEISHU_USER_ID"
feishu-bot calendar attendee list --calendar-id <calendar_id> --event-id <event_id>
feishu-bot calendar attendee delete --calendar-id <calendar_id> --event-id <event_id> --attendee-id <attendee_id>

Use --body-json for rooms, recurrence, reminders, conferencing, and complex attendee/resource payloads.

Video Meetings

feishu-bot vc reserve apply --end-time <unix_seconds> --owner-id <open_id> --topic "AI sync"
feishu-bot vc reserve get --reserve-id <reserve_id>
feishu-bot vc reserve active-meeting --reserve-id <reserve_id> --with-participants
feishu-bot vc reserve update --reserve-id <reserve_id> --end-time <unix_seconds>
feishu-bot vc reserve delete --reserve-id <reserve_id>
feishu-bot vc meeting get --meeting-id <meeting_id>
feishu-bot vc meeting list-by-no --meeting-no 123456789 --start-time <unix_seconds> --end-time <unix_seconds>
feishu-bot vc meeting invite --meeting-id <meeting_id> --user <open_id>
feishu-bot vc meeting set-host --meeting-id <meeting_id> --user-id <open_id>
feishu-bot vc meeting end --meeting-id <meeting_id>
feishu-bot vc recording get --meeting-id <meeting_id>
feishu-bot vc recording start --meeting-id <meeting_id> --timezone 8
feishu-bot vc recording stop --meeting-id <meeting_id>
feishu-bot vc recording set-permission --meeting-id <meeting_id> --user <open_id>
feishu-bot vc report daily --start-time <unix_seconds> --end-time <unix_seconds>
feishu-bot vc report top-user --start-time <unix_seconds> --end-time <unix_seconds> --limit 10 --order-by 1
feishu-bot vc room list --page-size 20
feishu-bot vc room get --room-id <room_id>
feishu-bot vc room mget --room-id <room_id>
feishu-bot vc room-level list --page-size 20

Reserve APIs can use tenant or user auth; tenant-token reserve creation needs --owner-id. Invite/end/recording start/stop/permission usually require FEISHU_USER_ACCESS_TOKEN and host or participant permission. Setting a host can require both vc:meeting and vc:meeting.participant:write.

Minutes

feishu-bot minutes search --query "周会" --page-size 20
feishu-bot minutes get --minute-token <minute_token_or_url>
feishu-bot minutes artifacts --minute-token <minute_token_or_url>
feishu-bot minutes media --minute-token <minute_token_or_url>
feishu-bot minutes transcript --minute-token <minute_token_or_url> --need-speaker --need-timestamp --file-format txt --output ./minute.txt

minute_token can be passed directly or copied as a full Feishu/Lark Minutes URL. search uses Feishu's user-token API, so set FEISHU_USER_ACCESS_TOKEN. Metadata, AI artifacts, media download URL, and transcript export use the normal app token path when the app has the required Minutes scopes and data access.

Search

feishu-bot search docs --query "飞书Bot" --page-size 10
feishu-bot search docs --query "方案" --doc-type DOCX --doc-type BITABLE --only-title
feishu-bot search message --query "上线" --chat-id oc_xxx --page-size 20
feishu-bot search source list --page-size 20
feishu-bot search schema create --file ./schema.json
feishu-bot search source create --name "AI 索引" --schema-id ai_schema --state 0
feishu-bot search item create --data-source-id <id> --id item_1 --title "标题" --url "https://example.com" --text "全文"

docs and message search visible content for the current user and therefore need FEISHU_USER_ACCESS_TOKEN. source/schema/item manage Feishu Open Search connector indexes with tenant_access_token; these are useful when an AI needs to push external project data into Feishu's unified search.

OKR

feishu-bot okr period list --page-size 20
feishu-bot okr period-rule list
feishu-bot okr user-okrs --user-id "$FEISHU_USER_ID" --offset 0 --limit 5
feishu-bot okr user-okrs --user-id "$FEISHU_USER_ID" --period-id <period_id> --offset 0 --limit 5
feishu-bot okr batch-get --okr-id <okr_id> --lang zh_cn

OKR commands use tenant_access_token and need scopes such as okr:okr.period:readonly, okr:okr:readonly, or okr:okr. period-rule and user OKR reads may also depend on the tenant's OKR edition and data visibility.

Attendance

feishu-bot attendance group list --page-size 20
feishu-bot attendance group get --group-id <group_id>
feishu-bot attendance shift list --page-size 20
feishu-bot attendance shift query --shift-name "早班"
feishu-bot attendance schedule query --user-id <employee_id> --from 20260501 --to 20260531
feishu-bot attendance task query --user-id <employee_id> --from 20260501 --to 20260531 --ignore-invalid-users
feishu-bot attendance flow query --user-id <employee_id> --from-ts 1760000000 --to-ts 1760086400
feishu-bot attendance stats query --user-id <employee_id> --operator-user-id <employee_id> --from 20260501 --to 20260531

Attendance commands use tenant_access_token. Group and shift reads need attendance:rule or attendance:rule:readonly; schedules, task results, flow records, and stats need attendance:task or attendance:task:readonly. Employee IDs default to employee_id; use --employee-type employee-no for work numbers. flow import and flow delete accept raw Feishu JSON for explicit write operations; flow delete accepts at most 10 record IDs.

Mail

feishu-bot mail message list --mailbox me --page-size 10
feishu-bot mail message get --mailbox me --message-id <message_id> --format metadata
feishu-bot mail folder list --mailbox me
feishu-bot mail settings send-as --mailbox me
feishu-bot mail settings accessible --mailbox me
feishu-bot mail contact list --mailbox me --page-size 20
feishu-bot mail message send --mailbox me --to user@example.com --subject "hello" --text "body"

--mailbox me uses FEISHU_USER_ACCESS_TOKEN. Explicit mailbox addresses can use tenant-token reads with --auth tenant, but the app must be granted Mail data resource permissions. Full mail bodies, subjects, addresses, and contact fields require separate Mail field scopes.

CoreHR

feishu-bot corehr department search --page-size 20 --field department_name --field code
feishu-bot corehr department get --department-id <id> --field department_name
feishu-bot corehr job list --page-size 20
feishu-bot corehr job get --job-id <id>
feishu-bot corehr job batch-get --job-id <id> --field job_name
feishu-bot corehr job-data query --employment-id <id> --page-size 20
feishu-bot corehr job-data get --job-data-id <id>
feishu-bot corehr person get --person-id <id>
feishu-bot corehr process list --modify-time-from <ms> --modify-time-to <ms> --page-size 20
feishu-bot corehr process get --process-id <id>

CoreHR commands use tenant_access_token. They require both Open Platform CoreHR scopes and Feishu People data-range grants. For complex filters, pass the official request body with --body-json, --file, or --stdin.

Helpdesk

feishu-bot helpdesk ticket list --page-size 20
feishu-bot helpdesk ticket get --ticket-id <ticket_id>
feishu-bot helpdesk ticket messages --ticket-id <ticket_id> --page-size 20
feishu-bot helpdesk service start --open-id <open_id> --human-service
feishu-bot helpdesk message send --receiver-id <open_id> --text "hello"
feishu-bot helpdesk faq categories --lang zh_cn
feishu-bot helpdesk faq list --search "登录" --page-size 20

Helpdesk commands use tenant_access_token plus the service-desk credential header X-Lark-Helpdesk-Authorization. Set FEISHU_HELPDESK_ID and FEISHU_HELPDESK_TOKEN from Helpdesk admin API credentials. Reads need helpdesk:all:readonly, service start needs helpdesk:helpdesk:access, and bot pushes need helpdesk:all. For future official payloads, pass native JSON with --body-json, --file, or --stdin.

Hire

feishu-bot hire job list --page-size 20
feishu-bot hire job detail --job-id <job_id>
feishu-bot hire job schemas --scenario 1
feishu-bot hire process list --page-size 50
feishu-bot hire talent list --keyword "张三" --page-size 10
feishu-bot hire talent get --talent-id <talent_id>
feishu-bot hire application list --talent-id <talent_id> --page-size 20
feishu-bot hire application detail --application-id <application_id> --option with_job --option with_talent
feishu-bot hire interview by-talent --talent-id <talent_id>
feishu-bot hire requirement schemas --page-size 20
feishu-bot hire metadata resume-sources --page-size 20
feishu-bot hire location query --location-type 1 --code CN_1

Hire commands use tenant_access_token and Feishu Hire data ranges. Start with read-only commands to discover real jobs, talents, applications, processes, and schema IDs. Writes are explicit: talent create and job open require typed flags or official JSON via --body-json, --file, or --stdin.

Wiki

feishu-bot wiki route-check
feishu-bot wiki route-check --write-probe
feishu-bot wiki route-check --write-probe --strict
feishu-bot wiki create-space --name "AI 知识库"
feishu-bot wiki spaces
feishu-bot wiki space --space-id <space_id>
feishu-bot wiki nodes --space-id <space_id>
feishu-bot doc create --writer official --title "AI 演示" --file ./demo.md --wiki --wiki-space-id <space_id> --wiki-fallback-ok
FEISHU_DOC_CREATE_WIKI_DEFAULT=true FEISHU_WIKI_SPACE_ID=<space_id> feishu-bot doc create --writer official --title "AI 演示" --file ./demo.md
feishu-bot wiki create-node --space-id <space_id> --title "AI 演示" --obj-type docx
feishu-bot wiki move-docs-to-wiki --space-id <space_id> --obj-type docx --obj-token <document_id>
feishu-bot wiki move-node --space-id <space_id> --node-token <node_token> --target-parent-token <parent_node_token>
feishu-bot wiki copy-node --space-id <space_id> --node-token <node_token> --target-space-id <space_id> --title "副本"
feishu-bot wiki update-title --space-id <space_id> --node-token <node_token> --title "新标题"
feishu-bot wiki node --token <node_token>
feishu-bot wiki task --task-id <task_id>
feishu-bot wiki member list --space-id <space_id>
feishu-bot wiki setting update --space-id <space_id> --create-setting admin_and_member
feishu-bot wiki search --query "关键字"

create-space and search use FEISHU_USER_ACCESS_TOKEN because Feishu's official APIs require user token there. Tenant-token Wiki calls only operate on spaces where the app or bot is already a member/admin. Set FEISHU_DOC_CREATE_WIKI_DEFAULT=true, FEISHU_WIKI_SPACE_ID, and optional FEISHU_WIKI_PARENT_NODE_TOKEN when AI dogfood documents should go to Wiki by default via feishu-bot doc create; use --no-wiki for an explicit one-off draft. Default Wiki publishing keeps the fallback docx and returns wiki_move_error when the move is blocked; pass --wiki-strict if the command should fail on a Wiki move error. Explicit one-off --wiki commands without the default route can pass --wiki-fallback-ok for the same fallback behavior. Run feishu-bot wiki route-check --json before claiming the default route is ready; it verifies the configured space plus read access for spaces, the target space, and child nodes without creating data. Run feishu-bot wiki route-check --write-probe --json before claiming future AI reports can all go through Wiki; it creates a small proof docx and attempts the same move_docs_to_wiki publish path used by doc create. Use --strict in automation when the run should exit non-zero unless route_ready is true. Only count Wiki publishing as successful when the command returns wiki_move or feishu-bot wiki node --token <wiki_node_token> can read the moved node. A created docx URL without wiki_move is only the fallback document. For automated dogfood using explicit --wiki, add --wiki-fallback-ok so the command still returns and sends the fallback docx while Wiki permissions are being granted.

Sheets

Create a spreadsheet from zero, then inspect sheets and write values:

feishu-bot sheet create --title "AI 数据表" --folder-token <folder_token>
feishu-bot sheet get --spreadsheet-token <token>
feishu-bot sheet sheets --spreadsheet-token <token>
feishu-bot sheet get-sheet --spreadsheet-token <token> --sheet-id <sheet_id>
feishu-bot sheet add-sheet --spreadsheet-token <token> --title "数据" --index 1
feishu-bot sheet update-sheet --spreadsheet-token <token> --sheet-id <sheet_id> --title "新标题" --frozen-row-count 1
feishu-bot sheet copy-sheet --spreadsheet-token <token> --sheet-id <sheet_id> --title "副本"
feishu-bot sheet delete-sheet --spreadsheet-token <token> --sheet-id <sheet_id>
feishu-bot sheet values get --spreadsheet-token <token> --range Sheet1!A1:C10
feishu-bot sheet values update --spreadsheet-token <token> --range Sheet1!A1:B2 --values-json '[[1,2],[3,4]]'
feishu-bot sheet values append --spreadsheet-token <token> --range Sheet1!A:B --values-json '[["new","row"]]'
feishu-bot sheet values prepend --spreadsheet-token <token> --range Sheet1!A:B --values-json '[["top","row"]]'
feishu-bot sheet merge --spreadsheet-token <token> --range Sheet1!A1:C1 --merge-type MERGE_ALL
feishu-bot sheet unmerge --spreadsheet-token <token> --range Sheet1!A1:C1
feishu-bot sheet style --spreadsheet-token <token> --range Sheet1!A1:C1 --bold true --back-color fff2cc --border-type FULL_BORDER
feishu-bot sheet style --spreadsheet-token <token> --range Sheet1!D2:D20 --formatter "0.00%" --h-align 1

Use merge, unmerge, and style when the AI needs to produce readable spreadsheet reports rather than raw cell values. style wraps Feishu styles_batch_update with typed flags for font, color, alignment, formatter, border, and clean operations while still accepting raw JSON for new fields.

Approval

feishu-bot approval definition get --approval-code <code>
feishu-bot approval definition subscribe --approval-code <code>
feishu-bot approval instance list --approval-code <code> --start-time <ms> --end-time <ms>
feishu-bot approval instance query --approval-code <code> --instance-status PENDING
feishu-bot approval instance get --instance-code <code>
feishu-bot approval instance create --body-json '{...}'
feishu-bot approval instance cancel --approval-code <code> --instance-code <code> --user-id <open_id>
feishu-bot approval task search --approval-code <code> --task-status PENDING
feishu-bot approval task approve --approval-code <code> --instance-code <code> --task-id <task_id> --user-id <open_id> --comment OK
feishu-bot approval task reject --approval-code <code> --instance-code <code> --task-id <task_id> --user-id <open_id> --comment "needs changes"
feishu-bot approval task transfer --approval-code <code> --instance-code <code> --task-id <task_id> --user-id <open_id> --transfer-user-id <open_id>
feishu-bot approval task add-sign --approval-code <code> --instance-code <code> --task-id <task_id> --user-id <open_id> --add-user-id <open_id> --add-sign-type 3
feishu-bot approval task rollback --task-id <task_id> --user-id <open_id> --task-def-key START --reason "revise"
feishu-bot approval external definition-get --approval-code <code>
feishu-bot approval external definition-create --file external-definition.json
feishu-bot approval external instance-sync --file external-instance.json
feishu-bot approval external instance-check --file external-check.json

Approval forms are schema-specific. Use definition get or OpenAPI Explorer to copy official form JSON and task IDs before creating instances or operating tasks.

Raw OpenAPI

Use feishu-bot api for official endpoints that do not have typed wrappers yet. It supports tenant/user token selection, JSON requests, binary downloads, and multipart uploads:

feishu-bot api get --path /bitable/v1/apps/<app_token>
feishu-bot api get --auth user --path /search/v2/data_sources
feishu-bot api post --path /task/v2/tasks --body-json '{"summary":"raw task"}'
feishu-bot api download --path /drive/v1/files/<file_token>/download --output ./file.bin
feishu-bot api multipart --path /im/v1/images --field image_type=message --file image=./image.png

Browser bridge

The browser commands call the local Playwright MCP extension bridge used by Codex:

feishu-bot browser ensure
feishu-bot browser tabs
feishu-bot browser drive

Name		Name	Last commit message	Last commit date
Latest commit History 124 Commits
.github		.github
docs		docs
private		private
scripts		scripts
skills/feishu-bot-office		skills/feishu-bot-office
src		src
.env.example		.env.example
.gitignore		.gitignore
CONTRIBUTING.md		CONTRIBUTING.md
Cargo.lock		Cargo.lock
Cargo.toml		Cargo.toml
LICENSE		LICENSE
README.md		README.md
SECURITY.md		SECURITY.md

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Repository files navigation

飞书Bot (`feishu-bot`)

Code layout

Quick checks

AI handoff

Messages

Chats

Directory

Notifications

Documents

Board / Mermaid

Base / Bitable

Tasks

Drive

Calendar

Video Meetings

Minutes

Search

OKR

Attendance

Mail

CoreHR

Helpdesk

Hire

Wiki

Sheets

Approval

Raw OpenAPI

Browser bridge

About

Uh oh!

Releases

Packages

Uh oh!

Contributors

Uh oh!

Languages

Folders and files

Latest commit

History

Repository files navigation

飞书Bot (feishu-bot)

Code layout

Quick checks

AI handoff

Messages

Chats

Directory

Notifications

Documents

Board / Mermaid

Base / Bitable

Tasks

Drive

Calendar

Video Meetings

Minutes

Search

OKR

Attendance

Mail

CoreHR

Helpdesk

Hire

Wiki

Sheets

Approval

Raw OpenAPI

Browser bridge

About

Topics

Resources

License

Contributing

Security policy

Uh oh!

Stars

Watchers

Forks

Releases

Packages 0

Uh oh!

Contributors

Uh oh!

Languages

飞书Bot (`feishu-bot`)

Packages