Felo Search on Doubao WorkDoubao Work

The deliverable is a file. The file still has to be right.

Doubao Work turns a request into a report, a deck, or a spreadsheet, on demand or on a schedule. Felo Search decides what goes into it: dated answers with the page behind each conclusion, so the finished file survives the first person who checks it.

Built for Doubao WorkWorks with scheduled tasksMarks what is unverified
Get an API keySee the platform guide
The command underneathfelo search "policy changes in the last 90 days"
  1. Bundlefelo-skills/felo-search
  2. RouteImport, or the Felo CLI
  3. Comes back asClaim, source, date
  4. On a scheduleDiffs against last week
Doubao Work logo
This skill runs on

Doubao Work

ByteDance's Doubao Work · agent work platform

Read the Doubao Work guide

Doubao Work turns a request into a finished file and can run it on a schedule. A Felo skill decides what goes into that file and keeps the sources behind every conclusion.

Runs on
Personal, team, and enterprise builds
Skills arrive by
Skill panel import, or the Felo CLI
Automation
Scheduled tasks and computer use
Add it on Doubao Work

Two ways in, depending on what your build exposes

Skill and connector availability differs between personal, team, and enterprise accounts, which is why there are two routes instead of one. Neither is better; one is simply available more often.

Path A

Import the skill bundle

The shorter route, when your build exposes a skill or connector panel. Personal, team, and enterprise builds differ here, so check what your own account shows before planning around it.

  1. Clone the repositorygit clone https://github.com/Felo-Inc/felo-skills.git. Every bundle is a folder with a SKILL.md, and you only need the one you are installing.
  2. Import the folderBring that folder in through the skill or connector panel so live search results with sources is available while the task runs, not only afterwards.
  3. Confirm it is callableAsk the agent which skills it can use before you build a scheduled task around one. An imported skill that never gets called is the usual failure here.
Path B

Drive the Felo CLI

The route that works on every build. Install it once, then either run it yourself during setup or let computer use invoke it as part of a task on your machine.

  1. Install and store the keynpm install -g felo-ai felo config set FELO_API_KEY your-api-key-here
  2. Run the command by hand oncefelo search "latest industry policy" # the same command the task will call
  3. Then fold it into the taskAsk the agent to call the CLI when it needs current information or a deliverable, so the same command runs inside the task instead of beside it.
What you get back

A claim list where every line can be defended in the meeting

A research report is only as good as the weakest line in it, so the output keeps the source and the date attached to each conclusion — and keeps the unverified material visibly separate instead of blending it in.

ConclusionSourceDate
The subsidy cap rises to 30 percent in Q4mof.gov.cn/notice/2026-083 weeks ago
Filing deadlines move to the 15thsamr.gov.cn/announcement/44713 weeks ago
The pilot extends to two more provincesndrc.gov.cn/news/pilot-round-36 weeks ago
The component tariff is unchangedcustoms.gov.cn/tariff-20269 days ago
Labelling may change in 2027draft rule, not in forceto confirm

Put the rule in the prompt rather than hoping for it: cite a source and a date for every conclusion, and write "to confirm" beside anything that is a draft, a rumour, or unsupported.

The routine

The weekly research job that gets better instead of longer

The point of scheduling this on Doubao Work is not to run more often, it is to make each run cheaper to read than the last.

Runs when

A recurring report — policy, market, competitive — where most weeks change nothing and one week changes everything.

  1. 01

    Run it once by hand and read the sources

    Before anything is scheduled. A recurring job that produces unchecked claims scales the mistake rather than the value.

  2. 02

    Ask for changes, not for a report

    Have the task report only what is new since the last run. The archive you already have is the baseline, and re-reporting it every week is how a useful digest becomes unread.

  3. 03

    Keep the "to confirm" column alive

    Draft rules, rumoured changes, and unsourced claims belong in their own bucket. When one of them firms up, that is itself the news.

  4. 04

    Diff against last week before you send it

    Ask what changed between runs. A report that contradicts its own last edition costs more trust than a quiet week does.

Leaves behindA short, dated list of what moved this week, with a separate bucket for anything unverified and the sources attached.

In practice

Two prompts that produce a defensible report

Both work whether the skill was imported or driven through the CLI, because the requirement lives in the request.

Type thisResearch this policy area for the last three months and produce a report. Cite a source and a date for every conclusion, and mark anything unverified as "to confirm".What comes back

A report whose every line can be traced, with the soft claims quarantined instead of mixed in.

Type thisTake the report from last week and produce only what changed since, with sources. If nothing changed, say so instead of padding.What comes back

A short delta, or a clear "no material change" that is itself worth knowing.

Where it stops

Three things to keep in mind before this runs unattended

Scheduling a research task raises the cost of a bad habit, so the limits matter more here than on a one-off run.

A schedule does not verify anything

Automation repeats a process; it does not improve one. If the manual run produced claims you could not check, the scheduled version produces those claims every week.

Sources go stale in different ways

A tariff page stays true for months; a news article is a snapshot. The dates are what let a reader tell which is which, so they belong in the output.

Import availability is a build question

If the skill panel is not visible on your account, that is an entitlement question rather than a setup mistake. Path B exists for exactly that case.

FAQ

What Doubao Work users ask about Felo Search

Does this replace the built-in research in Doubao Work?
It complements it. The platform is good at producing the deliverable; what a Felo skill adds is the sourcing discipline inside it — a link and a date per conclusion, and a visible bucket for what could not be checked.
Can computer use really run the CLI for me?
The CLI is a normal command-line tool, so a task that can operate your machine can call it. Installing it and storing the key first is what makes that reliable rather than hopeful.
What does the scheduled version cost?
One run per schedule, plus whatever the task does with the result. Nothing is fetched between runs, so a weekly job is one weekly job.
Where should the output live?
Wherever the team already reads it — a workspace document, a spreadsheet, a message. A report that only exists inside one task's history is not a deliverable.
More on Doubao Work

Every Felo skill for Doubao Work

Five skills written for Doubao Work. They follow the same pattern on this platform, so you can combine them in one session without re-explaining your setup.

Felo SearchLive web research with sources you can citeFelo SlidesGenerate a deck from a single promptFelo YouTube SubtitlingTranscripts from talks, demos, and reviewsFelo X SearchTrack posts and replies on XFelo SuperAgentLong-running workflows that keep their memory
Beyond ready-made skills

One API key. A full agent toolkit.

Start with a skill, then compose the exact models, live data, knowledge, and output tools your Doubao Work workflow needs — the same key every Felo skill already uses.

Models

Use cost-efficient LLM access when your agent needs custom reasoning or generation.

  • LLM API
  • Chat completions
  • Model access
Explore LLM API →

Live information

Give agents current web and social signals instead of relying on stale context.

  • Web Search
  • X Search
  • Web Fetch
Explore live data tools →

Knowledge and output

Ground work in reusable knowledge, then turn it into shareable deliverables.

  • LiveDoc
  • PPT generation
  • Mindmaps
Explore knowledge tools →

Creative tools

Extend the workflow from research to visual and office-ready results.

  • Image generation
  • Page tools
  • Office outputs
Explore platform tools →
Explore API Platform →Read the API docs →

Keep reading

Felo Search on its own pageThe skill, apart from any single platformThe Doubao Work guideDeliverables, scheduled tasks, and computer useFelo Slides for Doubao WorkWhen the report has to become a deck