Auto Study
ReviewAudited by ClawScan on May 10, 2026.
Overview
Auto Study is transparent about automating study websites, but it includes an unsafe WSL instruction to work outside the sandbox and can act through a logged-in browser profile.
Install only if you are comfortable with an agent controlling a dedicated Chrome profile for your study accounts. Do not use it for formal exams or where platform/institution rules prohibit automation. Review answers before any submission, keep all file operations inside the normal workspace/sandbox, and delete saved screenshots or records when finished.
Findings (5)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
The agent may try to run commands outside normal containment, making file writes and command side effects harder to limit or review.
This explicitly suggests bypassing the normal sandbox boundary for shell-based file operations, which is broader than the stated browser-study workflow.
If a socket error occurs while writing task artifacts, try using a shell loop outside the sandbox.
Remove or ignore the outside-sandbox workaround; require explicit user approval for any shell command and keep artifacts inside the intended workspace.
Actions may change answers or submit work on a logged-in learning platform if the user asks for it.
The skill gives the agent direct browser-control authority to change page state and potentially submit answers; this is central to the stated purpose and submission is separately restricted.
Interact with Chrome using CDP... Interact with the page according to the user's request, such as selecting, filling, or clearing answers, or clicking the submit button.
Use only where permitted, review answers before submission, and give explicit approval for each submit action.
Anyone or any agent using that profile may be able to access the same logged-in learning accounts and perform account actions.
The skill intentionally preserves login state in a persistent browser profile so the agent can act as the user on study platforms.
Reuse the same browser profile for the same site when login state matters. Always launch Chrome with the designated persistent profile.
Use a dedicated profile, avoid storing unrelated accounts in it, and clear or revoke sessions when no longer needed.
Study content, answers, scores, or personal learning-platform details may remain on disk after the task.
The workflow creates persistent local records containing screenshots, questions, answers, and potentially scores or identifying page details.
`record.md`; `images/full.png` for a full-page screenshot of the task; `images/q001.png`, `images/q002.png` for per-question screenshots
Review and delete stored artifacts when finished, especially on shared machines or sensitive coursework.
A user could unintentionally use the automation on a consequential quiz, assignment, or exam page if they do not verify the context first.
The skill relies on the user to identify formal exams or disallowed use rather than treating page labels as sufficient warning.
Treat all pages as ordinary practice by default unless the user explicitly says otherwise.
Stop and ask for confirmation when a page appears to be graded, formal, contest-based, or institutionally restricted.