The Agent Skills Directory

[SAFE]: No security issues were detected. The skill is entirely educational and provides templates for infrastructure automation within the Infrahub platform.
[PROMPT_INJECTION]: The instructions focus on technical implementation and do not contain any patterns intended to bypass AI safety filters or override system prompts.
[DATA_EXFILTRATION]: The skill uses the official infrahub-sdk to interact with its host platform. It does not attempt to access sensitive local files (e.g., SSH keys, environment variables) or send data to unauthorized external domains.
[REMOTE_CODE_EXECUTION]: There are no instances of downloading and executing scripts from remote sources. All code samples are local Python classes intended to run within an Infrahub environment.
[COMMAND_EXECUTION]: The skill mentions the infrahubctl CLI tool for testing generators, which is the standard management utility for the platform and does not pose an arbitrary command execution risk.
[INDIRECT_PROMPT_INJECTION]: The skill describes a data processing workflow where GraphQL query results are ingested by Python scripts.
Ingestion points: Data enters through the generate(self, data: dict) method in Python scripts located in the generators/ directory.
Boundary markers: No specific delimiters or boundary markers are used in the code examples to separate data from instructions.
Capability inventory: Generators have the capability to create, update, or delete infrastructure objects via self.client.create() and obj.save().
Sanitization: Examples use a clean_data helper to process dictionary structures, though it focuses on data unwrapping rather than security sanitization. Since this tool operates on structured infrastructure data within a managed environment, the risk is consistent with standard platform usage.

infrahub-generator-creator