The Agent Skills Directory

[PROMPT_INJECTION]: The skill defines a process for ingesting user-supplied trigger conditions and message payloads to create automation specifications, which constitutes an indirect prompt injection surface.
Ingestion points: User descriptions of event-driven automation (e.g., "When [condition], notify [message]") as described in the 'When To Use' and 'Procedure' sections of SKILL.md.
Boundary markers: Absent; the skill does not instruct the agent to use delimiters or ignore embedded instructions within the user-provided triggers or payloads.
Capability inventory: The skill utilizes the curl binary and is intended to interface with the Slack API.
Sanitization: The procedure lacks instructions for sanitizing, escaping, or validating external content before it is used to generate payloads or API calls.
[COMMAND_EXECUTION]: The skill requires the curl binary to be available on the PATH. While this is a standard tool for the skill's stated purpose of sending messages, its use to execute network requests based on user-defined payloads presents a risk of command argument injection if the agent is not careful during string interpolation.
[DATA_EXFILTRATION]: The skill relies on the presence of a SLACK_TOKEN environment variable. The combination of access to this sensitive credential and the ability to perform network operations via curl creates a potential for data exfiltration if the agent is manipulated by adversarial input in the trigger/payload definition phase.

magic-mouth