The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill installs the Membrane CLI tool from the public npm registry to facilitate communication with the QADeputy API.
Evidence: npm install -g @membranehq/cli@latest
[COMMAND_EXECUTION]: The skill utilizes several shell commands via the membrane CLI for authentication, connection management, and executing actions within the QADeputy environment.
Evidence: Commands such as membrane login, membrane connect, and membrane action run are used to interface with the platform.
[DATA_EXFILTRATION]: No evidence of unauthorized data exfiltration was found. The skill explicitly instructs the agent to avoid asking for user secrets (API keys/tokens), favoring Membrane's secure connection management which keeps credentials on the server side.
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface as it ingests and processes data from the external QADeputy API which could theoretically contain malicious instructions.
Ingestion points: Output from membrane action run (SKILL.md).
Boundary markers: Absent; the skill does not explicitly define delimiters for external data.
Capability inventory: Subprocess execution via membrane CLI across multiple commands (SKILL.md).
Sanitization: No explicit sanitization or validation of the external API response is documented.

qadeputy