The Agent Skills Directory

[SAFE]: The skill promotes secure authentication practices by explicitly instructing the use of non-production, least-privilege principals for all capacity probes.
[SAFE]: Strong operational boundaries are established, prohibiting the evasion of documented limits, unauthorized load testing, or the use of multiple tenants to bypass quotas.
[EXTERNAL_DOWNLOADS]: Fetches official integration contracts and support documentation from the Mindtickle website to establish an evidence-based capacity model.
[INDIRECT_PROMPT_INJECTION]: The skill processes data from external documentation and local configuration files, creating a theoretical injection surface.
Ingestion points: Contract data is retrieved via WebFetch from the official Mindtickle domain, and local workload profiles are accessed through Read and Glob tools.
Boundary markers: The instructions guide the agent to inventory specific, discrete fields (e.g., pagination, batch size, concurrency), which focuses the model on structured data extraction rather than arbitrary instruction following.
Capability inventory: The skill is restricted to Write and Edit for versioning policies and test fixtures; it does not utilize shell execution, subprocesses, or dynamic context injection.
Sanitization: The instructions emphasize modeling demand and defining conservative retry budgets based on extracted data, providing a logical constraint against processing instructions embedded within documentation text.

mindtickle-rate-limits