The Agent Skills Directory

[SAFE]: The skill uses a dedicated helper script (scripts/crawlora.sh) that manages the CRAWLORA_API_KEY securely. By using a restricted temporary configuration file (mktemp, chmod 600) and the curl --config flag, it ensures the API key is never exposed in process listings or shell history.
[SAFE]: The helper script implements a strict allowlist for API routes and validates input parameters. It uses regular expressions and case statements to prevent path traversal and command injection, while restricting HTTP methods to GET as required by the skill's functionality.
[INDIRECT_PROMPT_INJECTION]: The skill's primary function involves processing external TikTok data (profiles, posts, and comments), which is a common surface for indirect prompt injection attacks.
Ingestion points: Data retrieved via scripts/crawlora.sh and interpreted in SKILL.md workflows.
Boundary markers: The instructions do not define specific delimiters for untrustworthy content during analysis.
Capability inventory: The skill's active capabilities are restricted to performing authenticated GET requests to the vendor's API domain using curl within the scripts/crawlora.sh script.
Sanitization: The helper script sanitizes request paths and query parameters to prevent injection; however, the agent is responsible for safely parsing the returned JSON data.

tiktok-creator-research