The Agent Skills Directory

[PROMPT_INJECTION]: The skill is designed to ingest and process external specification files from the specs/ directory to generate code and tasks. This creates a surface for indirect prompt injection, where an attacker could embed malicious instructions within a specification file that the agent might inadvertently execute during the implementation phase.
Ingestion points: Processes Markdown specification files and implementation plans from specs/ and .specify/memory/plans/.
Boundary markers: No specific delimiters or safety warnings are implemented to differentiate between descriptive specifications and potential instructions to the agent.
Capability inventory: The skill allows for file system modifications, Git operations (including push), and the execution of the /speckit.implement tool.
Sanitization: No explicit validation or filtering of specification content is defined before it is used to drive the implementation workflow.
[COMMAND_EXECUTION]: The skill instructs the agent to execute various shell-based commands and platform-specific tools to manage the development lifecycle.
Evidence: The skill utilizes Git for branch management (git checkout -b, git commit, git push) and invokes specialized implementation and validation tools (/speckit.implement, /speckit.tasks, /stackshift.validate, /stackshift.review).

implement