The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a local PowerShell script located at .specify/scripts/powershell/check-prerequisites.ps1 during the initialization step. This script is used to verify the project environment and extract metadata, but it executes code residing within the local repository.
[DYNAMIC_EXECUTION]: The skill features an 'Extension Hooks' system that reads command names from a configuration file (.specify/extensions.yml). For hooks marked as mandatory, the skill explicitly instructs the agent to execute these commands using the EXECUTE_COMMAND directive. This allows the project configuration to dynamically trigger the execution of arbitrary slash commands without direct user initiation per command.
[INDIRECT_PROMPT_INJECTION]: The skill processes external data from .specify/extensions.yml to determine which commands to execute. If a user analyzes a repository containing a malicious extensions file (e.g., from a third-party pull request), the skill will attempt to execute the commands specified by the attacker.
[DATA_EXPOSURE]: The skill is designed to read and process multiple core project artifacts, including spec.md, plan.md, tasks.md, and .specify/memory/constitution.md. While its purpose is read-only analysis, users should ensure these files do not contain hardcoded secrets, as the content is loaded into the agent's context for modeling and analysis.

speckit-analyze