The Agent Skills Directory

[SAFE]: No signs of obfuscation, hardcoded credentials, or persistence mechanisms were detected within the skill instructions.
[EXTERNAL_DOWNLOADS]: The skill requires the installation of the @membranehq/cli package from the npm registry, which is the official tool from the vendor (membranedev).
[COMMAND_EXECUTION]: The skill utilizes shell commands via the Membrane CLI to manage authentication and interact with external data services.
[PROMPT_INJECTION]: A surface for indirect prompt injection exists because the skill processes results from external actions.
Ingestion points: Data from Thoughtful GPT enters the agent's context through the output field of the membrane action run command.
Boundary markers: There are no explicit instructions or delimiters provided to prevent the agent from executing instructions that might be contained within the external data.
Capability inventory: The agent can execute shell-based actions using membrane action run and can dynamically generate new actions via membrane action create in SKILL.md.
Sanitization: There is no evidence of sanitization or validation performed on the external data before it is interpreted by the agent.

thoughtful-gpt