The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses shell commands to initialize and manage the development environment, such as git init, git clone, and git submodule add. It also executes environment-specific commands for building and testing, including npm run build, npm test, and pytest. Additionally, it employs grep and dirname to identify its installation path within the platform's plugin directory.
[EXTERNAL_DOWNLOADS]: The skill retrieves external code and configuration by cloning repositories and adding git submodules based on URLs provided by the user via arguments or interactive prompts.
[INDIRECT_PROMPT_INJECTION]: The skill presents an attack surface for indirect prompt injection as it automatically ingests and interprets data from untrusted project files (e.g., package.json, pyproject.toml, README.md) to determine the technology stack and project state.
Ingestion points: Logic in references/stack-detection.md and SKILL.md (Steps 1 and 3.4) that parses project manifests and documentation.
Boundary markers: No specific delimiters or boundary markers are described to isolate data from instructions during parsing.
Capability inventory: The skill possesses capabilities to write files, execute shell commands, and invoke other agent skills based on the data it processes.
Sanitization: No explicit sanitization or validation of the content within the manifest files is described beyond standard format parsing.
[DYNAMIC_EXECUTION]: The skill dynamically generates configuration files (such as .sdd-devkit/settings.json) and agent instruction files (such as AGENTS.md) by interpolating detected project data into predefined templates.

arch-init