Scheduled tasks
Recurring work runs on a schedule without you opening the app. A research skill that returns consistent, structured output is what makes a recurring report worth reading.
Doubao Work is built around closing the loop: you describe the outcome, it plans the task, uses a domain skill library, and returns something usable — a deck with layout, a spreadsheet with formulas, a document with structure. It can also schedule that work and reach into your own machine to open files or operate a browser.
That raises the bar rather than lowering it. A generated report that cites nothing is still a liability when it reaches a client or a manager. A market number from training data is worse than no number, because it looks authoritative.
Felo skills give the pipeline an external evidence step: search with dated sources, fetch the actual pages, run a longer investigation when the question deserves it, and hand the findings back in time for the writing stage. The deck still comes from Doubao Work; the claims now trace to something.
Four capabilities that decide how you should install and call an external skill.
Recurring work runs on a schedule without you opening the app. A research skill that returns consistent, structured output is what makes a recurring report worth reading.
The agent can read and edit local files, open pages, and complete cross-page actions. It can also drive a local CLI when that is the shortest path to a result.
Built-in skills already cover investment, medical, creative, and e-commerce work. Custom or imported skills extend the same selection surface for your own sources.
Documents, knowledge, and conversation live with the workspace, which makes the handoff between research and delivery short.
Each one is written as the request, followed by what the Felo layer adds.
Every Monday morning, someone needs a summary of what competitors changed: pricing, packaging, launch posts, new case studies.
Every Monday at 8:00, check these competitors' pricing pages and public updates, then list what changed and what to watch this week.A recurring brief that flags what changed since the last run instead of re-describing the same pages, with links for every change.
A new product needs listing copy, a comparison table, and a short campaign brief — all grounded in what the category currently sells.
Research how the top products in this category are positioned and priced, then draft our listing copy and a launch plan.Copy and a plan built on observed category data, with the source list kept for the next iteration.
A clinician or researcher needs recent studies on a treatment question, sorted by evidence quality.
Find the multicentre studies on this treatment from the last five years and summarise efficacy, safety, and population, with the paper for each point.A literature summary where each line traces to a paper, ready to be reformatted into Doubao Work's document output.
An internal report exists; the meeting needs ten slides and a one-page summary.
Turn this report into a ten-slide deck and a one-page summary, keeping the key numbers and their sources.A slide draft and summary generated from the report, with the source notes preserved.
Two skills read the outside world, two produce the deliverable, and one generates the images that go with it.
Returns dated findings and source lists that a scheduled task can compare against its previous run.
Search this quarter's policy updates for the industry, list them by publication date, and give the source for each.A scheduled run should end in something you can forward. This turns the conclusions into a deck instead of leaving them as chat text.
Turn these findings into an eight-slide deck and put the data sources in the speaker notes.Plenty of Doubao tasks end in a picture — a product shot, a campaign visual, a logo. SuperAgent generates and revises those images in the same run.
Create a product hero image with a clean background and the packaging as the focus, plus two alternates in other styles.A webinar or a competitor's launch video becomes text the agent can quote in its report, which beats summarizing a video from the title.
Pull the transcript from this industry livestream and list the key numbers it mentions, with where each one came from.For a timed monitoring task: what customers and competitors posted today, in a form that can be diffed against yesterday's run.
Collect today's posts that mention our product, rank them by engagement, and give me the links.Doubao Work reaches these skills two ways: an imported bundle, or the Felo CLI on your machine. Pick the job first: research, extraction, slides, transcripts, social monitoring, or a long-running workflow.
Add on Doubao Work
Live web research with sources you can cite
Import felo-search through the skill or connector panelTurn pages into structured, reusable data
Import felo-web-fetch through the skill or connector panelAnswers grounded in your own documents
Import felo-livedoc through the skill or connector panelTranscripts from talks, demos, and reviews
Import felo-youtube-subtitling through the skill or connector panelTrack posts and replies on X
Import felo-x-search through the skill or connector panelGenerate a deck from a single prompt
Import felo-slides through the skill or connector panelTurn notes and URLs into a deck outline
Import felo-content-to-slides through the skill or connector panelLong-running workflows that keep their memory
Import felo-superagent through the skill or connector panelCapture pages as snapshots and Markdown
Import doc-screenshot-agent through the skill or connector panelDraft X posts from your own material
Import felo-twitter-writer through the skill or connector panelGenerate a landing page from a brief
Import felo-landingpage through the skill or connector panelApple product research before you buy
Import apple-buy-advisor through the skill or connector panelIf that panel is not in your build, install the Felo CLI once and let computer use call the skill during a task.
The reliable order is manual first, then scheduled: run the research by hand, check the sources, and only then let a task repeat it.
A timed task that produces unchecked claims scales the mistake, not the value.
Plenty of tasks end in something visual or in a recording nobody wants to rewatch.
Once the manual run is stable, scheduling is a one-line change.
Pick the path that matches your build and your appetite for local setup. Both end with the same result: a research step the agent can call during a task.
If your version of Doubao Work exposes a skill or connector import surface, use it. Skill bundles are folders with a SKILL.md, the same format Felo publishes.
git clone https://github.com/Felo-Inc/felo-skills.git
# import the felo-search folder through the skill or connector panelSkill and connector availability differs between personal, team, and enterprise builds. If the import entry is not visible on your account, use Path B.
Install the CLI once. Then either call it from a terminal during setup, or let Doubao Work's computer-use ability invoke it as part of a task on your machine.
npm install -g felo-ai
felo config set FELO_API_KEY your-api-key-here
felo search "latest industry policy"
felo slides "Q3 competitor review, 8 slides"Whether the skill is imported or driven through the CLI, tell the task what a good answer includes. Source requirements typed into the prompt are what keep the output reviewable.
Research this topic and produce a report:
- cite a source and a date for every conclusion
- mark anything unverified as "to confirm"
- finish with the source listRun the task once, read the sources, then move it to scheduled tasks. A recurring job that produces unchecked claims scales the mistake, not the value.
Schedule this research process for every Monday at 8:00
and diff the result against last week.Research skills make deliverables more defensible. These habits keep them that way.
A report without sources cannot be reviewed. Put the requirement in the recurring task definition so every run inherits it.
Generated analysis blends observation with advice. Ask for a summary that labels which sentences come from retrieved sources and which are inference.
The Felo CLI stores credentials in your user config. Do not paste keys into task text, where they become part of the conversation record.
Computer use and scheduled tasks can act without you watching. Put money, publishing, and messaging steps behind a manual approval.
Add one research skill, run a task you already do, and check whether the output is easier to review than the last one.
Felo AI is an independent skill publisher. Doubao Work is a ByteDance product.
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.
Use cost-efficient LLM access when your agent needs custom reasoning or generation.
Give agents current web and social signals instead of relying on stale context.
Ground work in reusable knowledge, then turn it into shareable deliverables.
Extend the workflow from research to visual and office-ready results.