The Agent Skills Directory

[COMMAND_EXECUTION] (SAFE): The skill utilizes curl for making HTTP requests and jq for processing JSON data. It employs bash -c to ensure environment variables are correctly passed through pipes, which is a documented workaround for specific agent behaviors and does not introduce security vulnerabilities in this context.
[DATA_EXFILTRATION] (SAFE): While the skill communicates with an external domain (api.openai.com), this is the primary and intended purpose of the skill. There is no evidence of unauthorized access to sensitive local files or exfiltration of private data.
[CREDENTIALS_UNSAFE] (SAFE): The skill correctly identifies OPENAI_API_KEY as a secret in the YAML frontmatter and uses placeholders in examples. No actual API keys or credentials are hardcoded.
[REMOTE_CODE_EXECUTION] (SAFE): The skill does not download or execute remote scripts or binary files. All commands are static and executed locally via curl and jq.

openai