The Agent Skills Directory

[COMMAND_EXECUTION]: Executes several shell commands including threat-designer, mmdc (Mermaid CLI), and git to analyze the repository, generate diagrams, and run the threat modeling engine.
[EXTERNAL_DOWNLOADS]: Instructs the user to install dependencies from official registries, specifically @mermaid-js/mermaid-cli from npm and the threat-designer tool via pip.
[DATA_EXFILTRATION]: Aggregates architectural and security metadata—including IAM policies, network configurations, and database schemas—to provide context for the threat modeling process. This information is processed by the threat-designer CLI, which returns a remote job ID, indicating interaction with an external analysis service to generate the threat model.
[PROMPT_INJECTION]: The skill ingests untrusted content from the local codebase (entrypoints, IaC files, schemas) to generate summaries and assumptions for the threat modeling tool. This creates a surface for indirect prompt injection (Severity: LOW).
Ingestion points: Section A2 (analysis of various repository files).
Boundary markers: Absent.
Capability inventory: threat-designer (CLI execution), mmdc (CLI execution), file system access, and network-interacting CLI.
Sanitization: Absent.

threat-designer