The Agent Skills Directory

[COMMAND_EXECUTION]: The skill dynamically executes commands based on the contents of .specify/extensions.yml. It parses this file for hooks.before_constitution and hooks.after_constitution and triggers execution via the EXECUTE_COMMAND directive. This allows for arbitrary command execution if an attacker can modify the project's extension configuration.
[INDIRECT_PROMPT_INJECTION]: The skill processes multiple untrusted data sources from the project repository to guide its behavior, creating a surface for indirect prompt injection.
Ingestion points: .specify/memory/constitution.md, .specify/extensions.yml, .specify/templates/plan-template.md, .specify/templates/spec-template.md, .specify/templates/tasks-template.md.
Boundary markers: Absent. The skill does not use specific delimiters or instructions to the model to ignore embedded prompts when reading these files.
Capability inventory: The agent has the ability to overwrite project files (.specify/memory/constitution.md) and execute system commands defined in the configuration.
Sanitization: Absent. Content from the templates and configuration is interpolated directly into the drafting logic and execution pipeline without validation or escaping.

speckit-constitution