Automations: run a real recurring job for a week and check it stays useful
メンテナーはふだん 2 日以内に返信
まだ誰も着手していません。
評価
調査の方向性
No code change is involved. Sign up at app.useagent.org, create one recurring Automation with clear instructions, and let it run at least five times. Done means a single report comment on this issue with links to your session threads, what worked, what broke, and what you automated and how often. File any bug as a separate issue and link it.
索引モデルが issue の本文から書いたものです。
説明
Check whether an automation stays useful over several runs.
Try this
- Pick something you'd want every day or week: a digest of a topic you follow, new releases of an open-source project you use, or new posts on a blog.
- Create it in Automations with clear instructions: what to check, what to write, how long, and "say nothing new if nothing changed".
- Turn it on and let it run at least five times.
- Read each result. Change the instructions once if something annoys you.
What to look for
- Results stay accurate and on topic, and links work.
- It doesn't repeat the same items run after run.
- It says "nothing new" when that's true, instead of padding.
- Failed runs are easy to spot.
- What would make you keep it running.
Your report
Post one comment here with:
- Links to your threads (
app.useagent.org/session/...) - What worked
- What broke or confused you, and what you expected instead
- Screenshots or a short screen recording
- What you automated and how often
Open a separate bug issue for anything broken, with the steps to reproduce it, and link it in your comment.
No need to claim; post your report as a comment.
General asks
- Do the steps yourself in the hosted app. Sign up with your email at app.useagent.org; you don't need an API key or a card to start. The link in the sign-up email works for 24 hours, so check spam if it's missing.
- Run everything by hand. Reports from runs done or simulated by another AI tool get closed.
- Use made-up details, test accounts and data you own. Don't point the agent at other people's accounts, inboxes or private data, and don't send real forms or messages to strangers.
- More than one person can report on the same issue, and every report helps. If someone already found what you found, add only what's different.
- Stuck? Comment here with a screenshot and we'll help.
Keep private details out of here
This page is public. Don't post your real email, phone number, address, API keys, or screenshots that show any of them; blur them first. Send anything private to [email protected] with a link to your comment here.
Using AI tools, and what gets no reply
- AI help is fine. Blind AI output is not. You must understand every line you submit, run it yourself, and show it working in the app with your own screenshots or recording.
- Pull requests that look generated and untested get closed without a reply and labeled
spam, which also means they don't count for Hacktoberfest. Signs we look for: no proof from a real run, code that ignores this repo's patterns or reinvents helpers that already exist, unrelated files changed, a description that only restates the issue, and failing checks nobody looked at. - One pull request per issue, by the person who claimed it. Opening pull requests on issues someone else has taken, or several near-identical pull requests for the same bounty, counts as farming and gets the same treatment.
- Bounties are paid only for the verified outcome the issue describes (for example, a workshop that actually happened, or an app that works on both systems), never for a write-up, a plan, or an unfinished attempt.
- When we ask a question in review, answer it in your own words. If you can't explain your change, we can't merge it.
- 主要言語
- TypeScript
- スター
- 371
- フォーク
- 78
- 平均マージ
- 2日 2時間
- マージ済み PR（30日）
- 44
環境構築
- Dockerfile・Docker Compose ファイルなし
- プルリクエストのテンプレートなし
- コントリビューションガイドを読む
はじめの一歩
- issue を最後まで読み、次にプロジェクトのコントリビューションガイドを読みます。
- 着手することを issue にコメントします — 二人が同じ作業をするのを防げます。
- リポジトリをフォークし、ブランチを切って変更します。
- issue 番号を参照したプルリクエストを送ります。
useagenthq/useagent のほかの issue
-
Split the Files panel view from its fetching and test each state対応中かも @sharmis-tp が 3 日前に担当しました。 オープンenhancement hacktoberfest help wanted
難易度 2/5 1〜3時間 初心者へのやさしさ 85/100
useagenthq/useagent#348 · コメント 1 件 ·
メンテナーはふだん 2 日以内に返信
-
Make the jump-to-message rail work on long threads and by keyboard対応中かも @sharmis-tp が 3 日前に担当しました。 オープンbug hacktoberfest help wanted
難易度 2/5 1〜3時間 初心者へのやさしさ 85/100
useagenthq/useagent#346 · コメント 1 件 ·
メンテナーはふだん 2 日以内に返信
-
Give the small buttons in a thread a 44px tap area on phones再び着手できるかも @xhbxx が 7 日前に担当しましたが、オープン中のプルリクエストはありません。 オープンaccessibility hacktoberfest help wanted
難易度 2/5 1〜3時間 初心者へのやさしさ 90/100
useagenthq/useagent#345 · コメント 1 件 ·
メンテナーはふだん 2 日以内に返信
-
Wrap long links and words in agent replies対応中かも @peterheo が 7 日前に担当しました。 オープンbug good first issue hacktoberfest
難易度 2/5 1〜3時間 初心者へのやさしさ 86/100
useagenthq/useagent#341 · コメント 2 件 · 担当者 1 名 ·
メンテナーはふだん 2 日以内に返信
-
enhancement good first issue hacktoberfest
難易度 2/5 1〜3時間 初心者へのやさしさ 86/100
useagenthq/useagent#340 ·
メンテナーはふだん 2 日以内に返信
useagenthq/useagent の issue をすべて見る
似ている issue
-
難易度 2/5 1〜3時間 初心者へのやさしさ 68/100
awesome-dsh-plugin/awesome-dsh-plugin#9184 ·
メンテナーはふだん 1 日以内に返信
-
RpcTest.makeClient: accept disableFatalDefects対応中かも @effect-bot が今日担当しました。 オープン
難易度 2/5 1〜3時間 初心者へのやさしさ 75/100
Effect-TS/effect#9893 · コメント 3 件 ·
メンテナーはふだん 1 日以内に返信
-
難易度 2/5 1〜3時間 初心者へのやさしさ 85/100
farion1231/cc-switch#8829 · コメント 1 件 ·
メンテナーはふだん 1 日以内に返信
-
[core-client] Remove unused hand-maintained transport helpers after the definition-table cutoverオープンdoc-audit enhancement
難易度 2/5 1〜3時間 初心者へのやさしさ 88/100
メンテナーはふだん 1 日以内に返信
-
難易度 2/5 1〜3時間 初心者へのやさしさ 85/100