Task Browser

Browse and filter all benchmark tasks.

134 tasks

ID ↑	Name ↕	Description	Difficulty ↕	Domain ↕	Factors	Mock Apps
1	skill-creation	Identify patterns from interaction history and create a SKILL from scratch	hard	Documents & Knowledge	B2
2	skill-supplementation	Supplement and update an existing SKILL based on interaction history	medium	Documents & Knowledge	B2
3	skill-conflict-resolution	Update existing SKILL based on interaction history: identify conflicting content, identify potentially misleading content in the old SKILL version, and update accordingly	easy	Documents & Knowledge	B2
4	skill-repository-curation	Organize existing SKILL library: merge redundant SKILLs, remove obsolete ones	easy	Documents & Knowledge	B2
5	skill-dependency-fix	After a user modifies a lower-level SKILL, identify the dependency relationships of higher-level SKILLs that call the lower-level SKILL, and update the higher-level SKILLs accordingly	easy	Documents & Knowledge	B2
6	email-writing	Help the user compose and send an email to a specified recipient	easy	Communication & Email		email
7	email-reply	Check if an expected email exists; if so compose and send a reply	easy	Communication & Email		email
8	flight-booking	Help the user book a flight for a specified date and route	medium	E-commerce & Daily Svcs		airline
9	flight-seat-selection	Help the user check in and select a seat for a specified booked flight; first retrieve the booking details from email, then select a seat on the airline website according to user requirements	medium	E-commerce & Daily Svcs	A1	airlineemail
10	flight-seat-selection-failed	Help the user check in and select a seat for a specified booked flight; first retrieve the booking details from email, then select a seat on the airline website; when user seat requirements cannot be met, find the best alternative	medium	E-commerce & Daily Svcs	A1B1	airlineemail
11	flight-cancel-claim	Check the status of a booked flight; if the flight is cancelled, help the user file a compensation claim according to the airline website announcement	hard	E-commerce & Daily Svcs	A1B1	airlineemail
12	flight-info-change-notice	Check the inbox for flight status change notifications from the airline; if found, check the calendar for affected plans and notify co-travelers	hard	Calendar & Task Mgmt	A1B1	airlineemailtodolist
13	baggage-tracking-application	Find a way to register lost luggage with the airline on its website	easy	E-commerce & Daily Svcs		airline
14	schedule-change-request	A schedule change requires the agent to find conflicting calendar entries, email relevant parties to inform them of the change, and explain the reason	medium	Calendar & Task Mgmt	A1	airlineemailtodolist
15	blog-site-from-scratch	Build a blog site from scratch: Astro frontend, Node.js backend, SQLite database; support regular user/editor/admin roles; support posting, keyword search, Markdown rendering, email registration/login, personal dashboard; editors can delete posts, admins can appoint editors, all pages support mutual navigation	easy	Coding & Software Dev
16	blog-site-completion-from-starter	Starting from a blog code skeleton, implement complete blog features: Astro frontend, Node.js backend, SQLite database; fill in user/editor/admin role permissions, search, Markdown rendering and other functionality	easy	Coding & Software Dev	A2
17	washer-shop	Buy a washing machine from Mosi Shop with a rating of at least 4.6 and portable	easy	E-commerce & Daily Svcs		shop
18	watch-shop	Buy a smartwatch from Mosi Shop with a rating of at least 4.6 and the lowest price	easy	E-commerce & Daily Svcs		shop
19	washer-change	The purchased item is not portable enough; exchange it for a washing machine rated at least 4.6 and portable, completing the return and re-ordering	easy	E-commerce & Daily Svcs		shop
20	info-change	Change shipping address to '4278 Maple View Drive, Sacramento, CA 95814, USA' and phone number to '12345678901'	easy	E-commerce & Daily Svcs		shop