The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill treats the story file as an authoritative implementation guide and is instructed to execute its tasks without deviation. Ingestion points include the story markdown file, sprint-status.yaml, and project-context.md. No explicit boundary markers or 'ignore embedded instructions' warnings are present. The agent has the capability to execute shell commands and modify files based on these untrusted inputs.
[COMMAND_EXECUTION]: The skill executes shell commands using uv run to call local scripts and standard tools like git. It also executes arbitrary test suites and linting tools discovered within the project structure, which can lead to the execution of code not controlled by the skill.
[DYNAMIC_EXECUTION]: In Step 10, the skill resolves a command string from project-controlled configuration files and is explicitly instructed to execute this value as a final terminal instruction. This allows for the execution of arbitrary shell commands defined in external project-level files.
[EXTERNAL_DOWNLOADS]: The use of the uv tool for running scripts may trigger the automatic download and installation of Python dependencies from external registries if they are not cached in the environment.

gds-dev-story