The Agent Skills Directory

[COMMAND_EXECUTION]: The skill is designed to automatically execute commands defined in the .specify/extensions.yml file using the EXECUTE_COMMAND: directive. This pattern allows for arbitrary shell command execution based on the content of a configuration file within the project directory, which is a high-risk capability.
[REMOTE_CODE_EXECUTION]: During the setup phase, the skill executes a PowerShell script located at .specify/scripts/powershell/setup-tasks.ps1. Execution of local scripts provided within a repository presents a high risk if the repository content has not been thoroughly audited.
[INDIRECT_PROMPT_INJECTION]: The skill acts as a processing engine for untrusted project data, creating a significant attack surface for indirect prompt injection.
Ingestion points: Project files including .specify/extensions.yml, plan.md, spec.md, data-model.md, contracts/, research.md, quickstart.md, and .specify/memory/constitution.md.
Boundary markers: No delimiters or instructions to ignore embedded commands are used when parsing these files.
Capability inventory: The skill has the ability to execute shell commands (EXECUTE_COMMAND), run PowerShell scripts, and perform file system read/write operations.
Sanitization: There is no evidence of input validation or sanitization for the data retrieved from the project files before it is used to generate tasks or trigger commands.

speckit-tasks