find a linear issue from the words you remember
stop asking the team who remembers that ticket. give the bot words from the problem, title or issue reference.
it searches linear and brings back matches, including archived work when you ask. use the separate team-listing action when you need one team’s backlog: putting a team, project or person’s name in this search only adds words.

search linear issues, in plain words
dig up the old investigation before repeating it. this search reads issue titles, descriptions and references. archived work stays out unless you ask for it.
you get 25 matches by default, with up to 50 in a batch. the bot can fetch more. team, project, owner, status and label filters need a different action; use list issues by team for a team’s issues.
a routine to copy
paste this into a chat with your bot
every monday at 11, search linear for "password reset", including archived issues.
- finish reading available results
- bring the issue titles and references here
- mark archived results separately
- flag loose matches
include the matching phrase beside loose matches. never edit, reopen or create an issue.- when
- mondays at 11:00
- needs
- linear connected with read only for this bot
what comes back looks like this (an example)
password reset matches
web-84: reset link opens a blank page; active
web-19: reset email delay; archived
web-84 is the closest match for the blank-page investigation.good jobs for it
- recover a support lead. find the ticket about a blank screen after sign-in.
- pick up an old investigation. include archives before reopening the same question.
- prepare for a support call. collect tickets mentioning the exact error phrase.
- find a half-remembered ticket. use words from its description.
before it runs
dusk bot counts Search Linear issues as a read: it looks and changes nothing. a bot you allow only to read can use it.
the permissions Linear asks for to run it: read.
how to start
- connect linear with read only. search needs no write access.
- give it the phrase. include the error words and say whether archives belong in the results.
- open the closest tickets. ask the bot to fetch remaining results before calling the list complete.
- schedule a topic worth watching. pause when the investigation ends.
limits worth knowing
- a team name is just another word here. use the separate team-listing action for a team boundary.
- bring the archives into an old-bug search. ask for them by name so earlier work appears.
- 50 is the largest batch. there may still be more matches to read.
the details it works from
| input | required | what it is |
|---|---|---|
query | yes | The search query text. |
after | no | Cursor for pagination. |
first | no | Number of issues to return per page. |
include_archived | no | Whether to include archived issues in search results. |
from composio's catalog for Linear (action LINEAR_SEARCH_ISSUES, read 12 november 20263). your bot fills these in from what you tell it.
try it in 2 minutes
- connect linear with read only
- quote a distinctive error phrase
- ask to include archived issues
- compare the matching title and description with your problem
is this the right search?
- you need words from a title, description or issue reference
- you chose whether to include archives
- you are not relying on team, owner, project, status or label filters
- you will read the matching issue before using it
mistakes people make
adding "ops team" and expecting a team filter
that only adds search words. use list issues by team when the team matters.
looking only at active tickets
include archives when you suspect someone solved this before.
questions
can it search issue comments?
not with this action. it checks the issue reference, title and description. reading comments takes a separate step.
what if i do not know the ticket number?
quote the problem instead. the bot can show candidates and ask which one you meant.
does including archives reopen anything?
no, it only finds them. archived issues stay as they are.
sources
drafted with ai help from the dusk bot docs and the sources listed on this page, then checked against them. if something here is wrong or out of date, write to [email protected].
change a linear issue without replacing the wrong details
a linear bot that spots missing context before planning
an ai assistant for product managers with scattered context
claude connectors: connect the job you actually need
find gmail messages from a half-remembered detail
an airtable assistant that learns the table before writing