The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes shell commands to verify the presence of development tools on the local system.
Evidence: Uses command -v within a bash block in SKILL.md to check for uvx, npx, and ccc binaries.
[EXTERNAL_DOWNLOADS]: The skill manages and recommends the installation of various tools from public package registries.
Evidence: The Pre-Configured MCP Stack and add sections in SKILL.md reference packages such as codebase-memory-mcp, serena, and @playwright/mcp to be executed via uvx and npx.
[PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface because it ingests data from local project files which could be manipulated to influence agent behavior.
Ingestion points: Reads .vscode/mcp.json and context/architecture.md (as described in SKILL.md).
Boundary markers: No explicit delimiters or instructions to ignore embedded commands are present when reading these files.
Capability inventory: The skill can execute shell commands (status check) and write to project configuration files (.vscode/mcp.json, .github/copilot-instructions.md).
Sanitization: Content from project files is interpolated into the agent's context without explicit validation or escaping.

mcp