The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes a shell script scripts/crawlora.sh to interface with the Crawlora API. The script follows security best practices by validating that only GET requests are permitted and enforcing a strict allowlist of authorized API paths. It prevents path traversal and parameter injection.
[CREDENTIALS_UNSAFE]: The skill requires an API key provided via the CRAWLORA_API_KEY environment variable. The helper script handles this securely by writing the key to a temporary file with restricted permissions (chmod 600) which is then passed to curl via a configuration file, ensuring the key does not appear in the system's process list.
[EXTERNAL_DOWNLOADS]: The skill makes network requests to api.crawlora.net. These requests are restricted to well-defined restaurant data endpoints and do not involve downloading or executing external code. The domain is the official API endpoint for the skill's author.
[INDIRECT_PROMPT_INJECTION]: The skill processes menu data (names, descriptions) fetched from external restaurant platforms. While this represents a data ingestion surface, the skill's capabilities are limited to data retrieval and formatting, with no high-risk actions (such as file writes or shell execution of data) that could be exploited by malicious content within the menu items.
Ingestion points: API responses from api.crawlora.net as referenced in SKILL.md and scripts/crawlora.sh.
Boundary markers: The instructions do not define specific delimiters for the external menu content.
Capability inventory: The skill uses a helper script for network requests but does not possess local file-write or arbitrary command execution capabilities.
Sanitization: The helper script sanitizes query parameters, and the skill's task is confined to structured data reporting.

restaurant-menu-benchmarking