Glasses to Social
Analysis
The skill is coherent, but it asks you to make a smart-glasses photo folder viewable to anyone with the link and can support periodic checks and social posting, so review privacy and account permissions carefully.
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.
Checks for instructions or behavior that redirect the agent, misuse tools, execute unexpected code, cascade across systems, exploit user trust, or continue outside the intended task.
Wait for "POST" confirmation or edits ... On approval, publish to configured platform (X/Twitter, etc.)
Publishing to a social platform is a high-impact account action, but the instructions explicitly require user approval before posting.
"schedule": {"kind": "cron", "expr": "*/15 * * * *", "tz": "UTC"}The skill documents periodic monitoring via cron. This is disclosed and user-directed, and the scheduled message asks the agent to draft rather than post automatically.
FOLDER_URL=$(jq -r '.googleDriveFolderUrl' "$CONFIG_FILE") ... gdown --folder "$FOLDER_URL" -O "$TEMP_DIR" --remaining-ok
The script depends on jq and gdown, while the registry metadata declares no required binaries and there is no install spec. The dependency use is purpose-aligned, but under-declared.
Checks whether tool use, credentials, dependencies, identity, account access, or inter-agent boundaries are broader than the stated purpose.
Create a shared Google Drive folder for glasses photos: ... Share with "Anyone with link can view"
The setup grants broad link-based access to a folder intended to contain smart-glasses photos, which may include private scenes, people, or text. The artifact does not provide a privacy warning or a narrower permission option.
Checks for exposed credentials, poisoned memory or context, unclear communication boundaries, or sensitive data that could leave the user's control.
For new photos, analyze with vision: ... Note any text, people, or notable elements
The workflow sends or exposes photo contents to a vision-capable model. This is core to the skill's purpose, but photos can contain sensitive people, locations, documents, or bystander information.
