The Agent Skills Directory

[COMMAND_EXECUTION]: Uses curl to facilitate message delivery to Slack and other webhook endpoints. This behavior is restricted to the skill's primary messaging function and occurs after a manual confirmation step.
[CREDENTIALS_UNSAFE]: Correctly identifies SLACK_TOKEN as a required environment variable for operation, avoiding the unsafe practice of hardcoding credentials within the instruction file.
[PROMPT_INJECTION]: The skill's instructions prioritize transparency and safety, specifically mandating that the agent must not impersonate users and must halt for confirmation before taking any live action.
[DATA_EXFILTRATION]: While the skill involves sending data to external services (Slack, webhooks, GitHub), this is done as a core feature. The inclusion of a mandatory human-in-the-loop approval process for every message mitigates the risk of unauthorized data transfer.

sending