Cardnews
WarnAudited by ClawScan on May 10, 2026.
Overview
The skill mostly matches its stated card-news purpose, but it can direct the agent to post publicly on Instagram and can silently install a Python package at runtime.
Use this skill only if you are comfortable reviewing every slide and caption before publishing. Do not allow it to upload to Instagram unless you explicitly confirm the target account and final post. Also note that running the JPG converter may install Pillow automatically; consider installing a pinned dependency yourself or editing the script to require confirmation.
Findings (3)
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 could post content publicly from an Instagram account when the user only expected image generation or upload preparation.
This tells the agent to use a browser workflow to publish a carousel post, but the artifacts do not require explicit final user confirmation, account selection, or draft-only handling.
Use the browser-based upload flow documented in TOOLS.md ("인스타그램 게시물 업로드" section). Upload all 5 JPG files as a carousel post.Make publishing optional, require an explicit final confirmation before upload/post, confirm the target Instagram account, and provide a draft/review mode.
The skill may act through whichever Instagram account is active in the browser, creating account-bound public content without a clearly declared permission boundary.
A browser-based Instagram upload depends on an authenticated Instagram account or session, but the metadata declares no credential or account boundary and the workflow does not specify which account is used.
### 5. Instagram Upload Use the browser-based upload flow documented in TOOLS.md
Declare the Instagram account/session requirement, ask the user to confirm the account before posting, and avoid using an ambient logged-in session without user approval.
Running the converter can unexpectedly modify the user's Python environment and execute newly downloaded dependency code.
If Pillow is missing, the converter silently installs an unpinned package from the Python package ecosystem at runtime, despite there being no install spec declaring this dependency.
subprocess.check_call([sys.executable, "-m", "pip", "install", "Pillow", "-q"])
Declare and pin Pillow as an installation dependency, avoid runtime auto-installation, or ask the user before installing packages.