The Agent Skills Directory

[DATA_EXPOSURE_AND_EXFILTRATION]: The skill correctly uses environment variables (e.g., $CO_API_KEY) for authentication rather than hardcoding secrets. It includes explicit instructions to exclude API keys and other sensitive data from logs, tickets, and error messages, which is a security best practice. Network operations are directed at well-known Cohere service domains.
[COMMAND_EXECUTION]: The skill uses curl for status checks and API connectivity verification, and npm/pip for version auditing. These tools are scoped to the primary purpose of the skill (diagnosing API issues) and are used in a non-malicious, diagnostic manner.
[INDIRECT_PROMPT_INJECTION]: While the skill is intended to ingest and analyze external data (error logs), it mitigates risks by instructing the agent to apply the 'smallest safe recovery path' and to redact sensitive correlation data before outputting results.

cohere-common-errors