The Agent Skills Directory

[SAFE]: The skill facilitates communication with the official LINE Messaging API via https://api.line.me. All network requests are directed to this legitimate service domain for bot management and messaging.
[COMMAND_EXECUTION]: Uses curl and jq for API interaction. The commands are structured to use bash -c to ensure environment variables like LINE_TOKEN are correctly passed within the execution context, which is a standard operational practice for shell-based agent skills.
[CREDENTIALS_UNSAFE]: Authentication is managed via the LINE_TOKEN environment variable, which is properly declared in the vm0_secrets metadata. There are no hardcoded secrets or credentials present in the skill's instructions or code snippets.
[DATA_EXFILTRATION]: No unauthorized data transmission was identified. The skill only sends data provided by the user or agent to the official LINE API as required for the messaging functionality.
[INDIRECT_PROMPT_INJECTION]: The skill processes external identifiers like USER_ID and REPLY_TOKEN. While these are ingestion points for untrusted data, the usage is consistent with the primary purpose of the skill, and commands are structured with quoting to mitigate basic shell injection risks.

line