The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests and processes external codebases which could contain malicious instructions designed to influence the agent's behavior during the modernization process.
Ingestion points: Legacy source code provided via Git URL or local path (legacy/$system).
Boundary markers: None specified for the analysis or processing of the ingested code in SKILL.md.
Capability inventory: The skill can execute file writes (write_to_file), clone repositories (git clone), and run modernization-specific commands (/code-modernization:*).
Sanitization: No explicit validation or sanitization of the legacy code content is described before processing.
[DYNAMIC_EXECUTION]: The skill dynamically reads and executes instructions from a local template file located at ${ANTIGRAVITY_PLUGIN_ROOT}/commands/modernize-preflight.md. While this is a common platform pattern, it involves executing instructions defined outside the primary skill file.
[COMMAND_EXECUTION]: The skill invokes several external commands, such as /code-modernization:modernize-status, /code-modernization:modernize-preflight, and scc, to analyze the project state and initiate the modernization workflow.
[EXTERNAL_DOWNLOADS]: The skill performs a git clone operation from a user-provided URL into the local legacy/ directory to obtain the source code for modernization.

modernize