The Agent Skills Directory

[REMOTE_CODE_EXECUTION] (CRITICAL): The Quick Install instructions contain a piped remote execution pattern: bash -c "$(curl -sL https://raw.githubusercontent.com/digitaladaption/cron-creator/main/install.sh)". This executes a remote shell script from an untrusted GitHub repository directly on the host machine without verification.
[EXTERNAL_DOWNLOADS] (HIGH): The manual installation instructions recommend git clone from github.com/digitaladaption, which is not a trusted organization per the [TRUST-SCOPE-RULE]. This involves downloading and running unverified code from an external source.
[COMMAND_EXECUTION] (HIGH): The skill documentation states that it configures tools.exec.host=gateway and executes clawdbot gateway restart. This modification of the execution context grants the skill broad permissions to execute commands on the gateway host, which could lead to privilege escalation.
[PROMPT_INJECTION] (LOW): The skill is highly vulnerable to Indirect Prompt Injection (Category 8) because it converts natural language input directly into system commands (clawdbot cron add).
Ingestion points: User chat messages (e.g., "Remind me to...") provided in SKILL.md.
Boundary markers: Absent; there are no delimiters or 'ignore' warnings for the natural language parser.
Capability inventory: The skill uses scripts/cron_creator.py to generate and likely execute CLI commands, and can restart the system gateway.
Sanitization: Absent; the documentation does not describe any escaping or validation logic for the input strings used in the cron command generation.

cron-creator