E
Ejk216/answer-entry
A Claude skill for working online homework — WeBWorK, Top Hat, and D2L quizzes: solving, entering answers,…
no SECURITY.md
View on GitHub (unreachable when checked on 2026-11-12)This link returned 404 on 2026-11-12
- Star count
- 0 ★ · Ejk216/answer-entry · read from the GitHub search API · (checked 2026-11-03 20:00:43)
- Source
- read directly from its GitHub page · type: open-source project
- Harvest history
- first seen 2026-10-22 · last seen 2026-11-14 · seen 27× · listed by 1 sources
- Score
- score 103 · 1 sources · repo
FAQ (answers come from our records)
Who is it for?
We filed it under these tags: 开源项目. The tags come from the entry itself, not a taxonomy we imposed.
Who maintains it?
repository
Ejk216/answer-entry;first catalogued 2026-10-22;last seen 2026-11-14。There is no official certification or quality rating here — we record sources and dates only.# answer-entry
以下摘录自该仓库的 README(片段) · 在 GitHub 读完整 README →
A Claude skill for working online homework — reading the questions, solving
them, and entering answers without burning attempts.
Covers WeBWorK, Top Hat, and D2L (Brightspace) quizzes.
Course-agnostic: nothing in it is tied to one school, server, or semester.
Each platform breaks in its own way, and the skill encodes what actually works
on each:
WeBWorK typed boxes are MathQuill live-math editors, not text inputs.
They rewrite input as you type and silently ignore programmatic fill — the
obvious approach submits a blank answer and spends a graded attempt telling
you nothing. Dropdowns and radios need the opposite method.
Top Hat hides correctness entirely, runs one question per page, and its
drag-and-drop needs the keyboard rather than the mouse.
D2L puts the quiz in a cross-origin iframe where text extraction returns
nothing and scrolling doesn't work — but the whole quiz is reachable from
JavaScript in a single pass.
Plus the parts that aren't platform-specific: a self-verification protocol for
when Claude is solving rather than transcribing, and the traps in common
problem types.
## Layout
SKILL.md — router: setup, modes, stakes, pipeline, speed and recovery rules
references/webwork.md, references/tophat.md, references/d2l.md — entry
mechanics per platform (read only the one you're on)
references/solving.md — self-verification and solution procedures
(Solve/Check modes only; Enter mode skips it)
harness/ — scripted WeBWorK read/enter (Playwright, persistent profile).
Faster than hand-driven entry with programmatic read-back before submit;
see harness/README.md. Optional — the skill works without it.
## Install
claude.ai — download this repo as a ZIP (green Code button →
Download ZIP), then Settings → Capabilities → Skills → upload it.
Claude Code / desktop — clone into your skills directory:
bash
git clone https://github.com/Ejk216/answer-entry.git \
~/.claude/skills/answer-entry
The file needs to land at ~/.claude/skills/answer-entry/SKILL.md.
Start a new session and Claude picks it up.
Updating: re-upload the ZIP after pulling, so the installed copy doesn't go
stale — compare the Version: line at the top of SKILL.md against the repo
if in doubt.
## What you need
A browser Claude can drive. Either works — no measured speed difference
between them, and they keep separate logins:
Claude in Chrome — the extension, driving your own Chrome
the built-in browser in the Claude desktop app — nothing to install
Expect one permission stop. Claude in Chrome gates access per site, so the
first time it opens your LMS, Top Hat or WeBWorK host you'll be asked to allow
that site. That's normal and one-time — allow it and tell Claude to continue.
You sign in. Claude never asks for credentials. If a page comes back as a
login screen, take the browser, sign in, and say continue.
Model and effort. Low effort is right for almost all of this — the clock is
dominated by browser round-trips, not thinking, and the skill carries the
procedure. A mid-tier model handles routine sets fine. Save a stronger model
them, and entering answers without burning attempts.
Covers WeBWorK, Top Hat, and D2L (Brightspace) quizzes.
Course-agnostic: nothing in it is tied to one school, server, or semester.
Each platform breaks in its own way, and the skill encodes what actually works
on each:
WeBWorK typed boxes are MathQuill live-math editors, not text inputs.
They rewrite input as you type and silently ignore programmatic fill — the
obvious approach submits a blank answer and spends a graded attempt telling
you nothing. Dropdowns and radios need the opposite method.
Top Hat hides correctness entirely, runs one question per page, and its
drag-and-drop needs the keyboard rather than the mouse.
D2L puts the quiz in a cross-origin iframe where text extraction returns
nothing and scrolling doesn't work — but the whole quiz is reachable from
JavaScript in a single pass.
Plus the parts that aren't platform-specific: a self-verification protocol for
when Claude is solving rather than transcribing, and the traps in common
problem types.
## Layout
SKILL.md — router: setup, modes, stakes, pipeline, speed and recovery rules
references/webwork.md, references/tophat.md, references/d2l.md — entry
mechanics per platform (read only the one you're on)
references/solving.md — self-verification and solution procedures
(Solve/Check modes only; Enter mode skips it)
harness/ — scripted WeBWorK read/enter (Playwright, persistent profile).
Faster than hand-driven entry with programmatic read-back before submit;
see harness/README.md. Optional — the skill works without it.
## Install
claude.ai — download this repo as a ZIP (green Code button →
Download ZIP), then Settings → Capabilities → Skills → upload it.
Claude Code / desktop — clone into your skills directory:
bash
git clone https://github.com/Ejk216/answer-entry.git \
~/.claude/skills/answer-entry
The file needs to land at ~/.claude/skills/answer-entry/SKILL.md.
Start a new session and Claude picks it up.
Updating: re-upload the ZIP after pulling, so the installed copy doesn't go
stale — compare the Version: line at the top of SKILL.md against the repo
if in doubt.
## What you need
A browser Claude can drive. Either works — no measured speed difference
between them, and they keep separate logins:
Claude in Chrome — the extension, driving your own Chrome
the built-in browser in the Claude desktop app — nothing to install
Expect one permission stop. Claude in Chrome gates access per site, so the
first time it opens your LMS, Top Hat or WeBWorK host you'll be asked to allow
that site. That's normal and one-time — allow it and tell Claude to continue.
You sign in. Claude never asks for credentials. If a page comes back as a
login screen, take the browser, sign in, and say continue.
Model and effort. Low effort is right for almost all of this — the clock is
dominated by browser round-trips, not thinking, and the skill carries the
procedure. A mid-tier model handles routine sets fine. Save a stronger model
Listed badge (optional)
Copy this line into the README if you want the listing mark. Optional, and it does not affect being listed.
[](https://anclum.com/en/s/ejk216-answer-entry/)
- Source
- github · https://github.com/Ejk216/answer-entry
- Browse by field
- Open source
- added
- today
- Security policy
- no
SECURITY.mdat the repository root (which does not mean it is unsafe)
More like this
Filed under
Every tag opens the full list for that filter. Counts are the live catalogue figures.