The Agent Skills Directory

[SAFE] (SAFE): No malicious behavior, prompt injection, or obfuscation was detected. The skill uses standard libraries for its intended purpose of Slack workspace automation.
[EXTERNAL_DOWNLOADS] (LOW): The documentation references official packages (slack-bolt, slack-sdk) and common utilities (python-dotenv, ngrok). These are from trusted sources and are standard for this type of development.
[CREDENTIALS_UNSAFE] (SAFE): The code examples correctly demonstrate accessing sensitive tokens (SLACK_BOT_TOKEN, etc.) via environment variables rather than hardcoding them.
[COMMAND_EXECUTION] (SAFE): The README includes instructions for the user to run pip install and curl for setup and testing. These are transparent and expected for a technical guide.
[INDIRECT_PROMPT_INJECTION] (LOW):
Ingestion points: The skill processes untrusted data from Slack users through message listeners and slash commands (e.g., app.message("hello")).
Boundary markers: No specific boundary markers or sanitization are shown in the basic examples.
Capability inventory: The skill has the capability to send and update messages in Slack channels.
Sanitization: The basic examples do not demonstrate input sanitization, which is a common pattern in introductory documentation but constitutes a potential surface for injection if the bot outputs user content directly into sensitive contexts.

slack-api