The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to process data from external sources such as Jira tickets and Slack messages, which may contain untrusted content.
Ingestion points: The jira and slack tools fetch attachment data and conversation history from remote Jira issues and Slack channels as described in SKILL.md.
Boundary markers: There are no specific instructions regarding the use of delimiters or sanitization steps when the agent processes this external content.
Capability inventory: The skill has access to bash commands to execute Jira/Slack CLI tools and the GitHub CLI (SKILL.md).
Sanitization: The provided instructions do not detail any filtering or escaping mechanisms for external data before it is handled by the agent.
[EXTERNAL_DOWNLOADS]: The documentation references an external Node.js package for tool installation and a GitHub repository for reporting issues.
Mentions installation via @aabuhijleh/courier and bug reporting at aabuhijleh/abed-hub. These resources belong to the identified skill author and represent standard vendor-owned dependencies.
[DATA_EXFILTRATION]: The skill accesses a sensitive local path to retrieve configuration and credentials necessary for its primary function.
The instructions direct the agent to read API tokens and base URLs from ~/.config/abed-hub/courier/config.json as noted in jira-adf.md. The skill implements best practices by requesting the user to handle credential creation, using restricted file permissions (mode 600), and explicitly warning the agent to avoid printing tokens to standard output.

courier