The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill installs a skill bundle from the author's repository using npx skills@latest add milad-alizadeh/argo and executes a synchronization script (hooks-sync.mjs) from a shallow-cloned version of the same repository.
[EXTERNAL_DOWNLOADS]: Fetches configuration templates and filter files directly from the author's GitHub repository (raw.githubusercontent.com/milad-alizadeh/argo) using curl during the initialization process.
[COMMAND_EXECUTION]: Employs standard CLI tools including git, gh, npx, jq, and node to manage repository state, fetch API data, and process local configuration files.
[DYNAMIC_EXECUTION]: Utilizes node to run an external script (bin/hooks-sync.mjs) found within a cloned repository to generate project-specific settings for Claude Code and Codex harnesses.
[INDIRECT_PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface by ingesting external, potentially untrusted data (such as project labels from gh label list) and resolving them into markdown templates (templates/labels.md) that guide agent behavior.
Ingestion points: External project labels retrieved via gh label list (SKILL.md, Phase 3).
Boundary markers: No specific boundary markers or "ignore" instructions are present in the documentation templates where labels are resolved.
Capability inventory: Includes file system write access, git checkouts, and shell command execution via node and npx.
Sanitization: The skill does not explicitly describe sanitization or validation steps for the label strings before they are interpolated into the documentation templates.

setup-argo-skills