The Agent Skills Directory

[PROMPT_INJECTION]: The skill creates an attack surface for indirect prompt injection by ingesting and acting upon untrusted data from external sources like GitHub issues, Package Contracts, and PRD documents.
Ingestion points: Reads delivery unit claims, GitHub Package Contracts, and PRD children from external repositories.
Boundary markers: The instructions lack explicit delimiters or "ignore embedded instructions" warnings to separate untrusted external content from the agent's core instructions.
Capability inventory: The skill allows file system modification via Git, execution of test suites, and network interaction with GitHub.
Sanitization: There is no mention of sanitization or validation of the fetched external text before it is processed by the agent.
[COMMAND_EXECUTION]: The skill performs shell-based operations for environment management and verification. This includes Git commands (e.g., git status, worktree isolation) and the execution of project-specific test runners and typecheckers. These capabilities are necessary for the skill's implementation purpose but involve executing code determined by the local environment and ingested specifications.

and-implement