The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute a local script spec-superflow.mjs using node. This script accepts arguments such as --verification-command and --test-cmd, which are subsequently executed as shell commands. This provides a direct path for executing arbitrary code on the host system depending on the values passed to these parameters.
[INDIRECT_PROMPT_INJECTION]: The skill determines its actions, including which verification commands to run, based on the contents of the target directory (<change-dir>). If an attacker provides a malicious project structure containing deceptive verification instructions, the agent might execute harmful commands.
Ingestion points: Project configuration and files located within the <change-dir> or <dir> paths.
Boundary markers: The skill does not define specific delimiters or instructions to ignore embedded commands within the processed project files.
Capability inventory: The skill can execute the spec-superflow.mjs script, which in turn can spawn shell processes for testing and verification.
Sanitization: No sanitization or validation logic is specified for the commands passed to the --verification-command or --test-cmd flags.

release-archivist