The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses shell commands for project environment management, including directory creation (mkdir -p), modifying file permissions to enable script execution (chmod +x), and computing file hashes using git hash-object. The configuration allows broad Bash tool usage.
[EXTERNAL_DOWNLOADS]: The skill retrieves configuration rules, scripts, and hooks from a local sd0x-dev-flow plugin installation. It discovers these resources using glob patterns to search system plugin directories and project dependencies.
[REMOTE_CODE_EXECUTION]: The skill installs shell scripts and JavaScript files into the project and configures the agent's settings.json to execute these files automatically as event-based hooks (e.g., pre-tool usage or session start).
[PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface by ingesting data from manifest files and interpolating it into project documentation.
Ingestion points: Project manifests including package.json, pyproject.toml, Cargo.toml, go.mod, build.gradle, pom.xml, and Gemfile.
Boundary markers: Absent; values are directly replaced in template files.
Capability inventory: Broad shell access via Bash, file creation via Write, and configuration editing via Edit.
Sanitization: Absent; the skill does not explicitly validate or escape data read from manifests before interpolation.

project-setup