The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes local shell scripts located in the .specify/scripts/bash/ directory. Specifically, it calls setup-plan.sh --json to initialize environment variables and update-agent-context.sh to modify agent-specific context files. These actions are standard for the skill's intended workflow but involve executing code from the repository.
[PROMPT_INJECTION]: The workflow processes content from specs/<feature>/spec.md and user-provided constraints, creating an indirect prompt injection surface.
Ingestion points: Reads and parses specs/<feature>/spec.md and user inputs.
Boundary markers: None identified; the instructions do not specify delimiters or warnings to ignore embedded instructions in the ingested files.
Capability inventory: Includes execution of local bash scripts and the ability to write or modify various files in the repository (e.g., research.md, data-model.md, and agent context files).
Sanitization: Not mentioned; the skill does not appear to validate or sanitize the content of the specification files before using them to generate design artifacts.

speckit-plan