The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructions direct the AI agent to execute multiple shell commands using the agent-kernel CLI tool, including architecture doctor, architecture discover, architecture reuse, and architecture check. These commands are used to analyze the codebase and enforce architectural rules.
[EXTERNAL_DOWNLOADS]: A provided GitHub Actions template (templates/github-actions.yml) includes a command to install the @mamdouh-aboammar/agent-kernel package from the NPM registry. This package appears to be the core toolset associated with the skill author.
[PROMPT_INJECTION]: The skill uses agent-kernel architecture discover . to ingest data from the local repository. This represents a surface for indirect prompt injection where instructions embedded in the codebase could attempt to influence the agent's analysis, although the tool is documented to perform static analysis without executing the source files.
Ingestion points: Files scanned by agent-kernel architecture discover . across the repository.
Boundary markers: The instructions do not explicitly provide boundary delimiters for the agent when reviewing tool output from the scanned code.
Capability inventory: The agent can execute the agent-kernel tool suite and perform file reads/writes within the defined contract scope.
Sanitization: The tool documentation (references/threat-model.md) outlines mitigations such as normalized paths and regex-based extraction to prevent code execution during analysis.

architecture-guardian