The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process documentation files such as markdown, YAML, SQL, and JSON from external sources (GitHub repositories or local directories). This content is then used to define the logic for writing analytics SQL. If an attacker can influence these documentation files, they could inject instructions that lead the agent to write incorrect, malicious, or data-leaking SQL queries.
Ingestion points: Remote repository contents fetched via gh api and local files in data_dictionary/ or models/ directories.
Boundary markers: None identified. The skill does not instruct the agent to ignore or delimit instructions found within the documentation files.
Capability inventory: The agent has the ability to read files, execute gh CLI commands, and generate SQL code based on the findings.
Sanitization: No sanitization or validation of the fetched documentation content is described.
[COMMAND_EXECUTION]: The skill provides templates for shell commands using the GitHub CLI (gh) to interact with the GitHub API. These commands are used to list directory contents and read files. While these are standard development operations, they involve executing shell commands with interpolated repository and file paths.

reading-data-dict