The Agent Skills Directory

[COMMAND_EXECUTION]: The skill is configured to use the shell MCP server, enabling the execution of shell commands via mcp__shell__run_short_command and mcp__shell__start_command as defined in agents/openai.yaml.
[EXTERNAL_DOWNLOADS]: The skill has the capability to fetch content from external URLs using mcp__docs_web__fetch_url and clone remote Git repositories using mcp__git__clone_git_repository.
[PROMPT_INJECTION]: The skill exhibits a vulnerability surface for indirect prompt injection where data from external sources could influence agent behavior.
Ingestion points: Untrusted external data enters the context via mcp__docs_web__search_web and mcp__docs_web__fetch_url tools.
Boundary markers: The instructions in SKILL.md and the prompt templates do not include delimiters or instructions to ignore embedded commands within fetched content.
Capability inventory: The skill possesses extensive tools, including shell execution (mcp__shell__run_short_command), file modification (mcp__workspace__write_file), and remote repository interaction (mcp__git__push_git_branch).
Sanitization: The skill lacks logic to sanitize, escape, or validate data retrieved from external URLs before it is processed by the agent.

frontend-delivery