The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses child_process.spawnSync to execute git commands and local Node.js scripts for project initialization and delta scanning.
[PROMPT_INJECTION]: The skill facilitates indirect prompt injection by reading untrusted data from repository files (READMEs, documentation, and configuration) and incorporating it into the agent's context.
Ingestion points: scan_midway_repo.js reads README files and package scripts from the target repository.
Boundary markers: No delimiters or protective instructions are used to distinguish repository content from system instructions in the generated output.
Capability inventory: The skill scripts have access to the file system and can spawn subprocesses via Git and Node.js.
Sanitization: The skill does not sanitize or filter content read from the repository files before providing it to the model.

repo-intake