The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes specifications and tickets from an external issue tracker to guide its code implementation workflow. This creates a surface for indirect prompt injection where a malicious actor could embed instructions within a ticket to manipulate the agent or its subagents.
Ingestion points: The skill reads specifications and tickets from a provided issue tracker (mentioned in Step 1).
Boundary markers: There are no explicit instructions to treat the content of tickets as data or to use delimiters to prevent command leakage into the agent context.
Capability inventory: The skill possesses significant capabilities, including file system access (creating worktrees, saving markdown notes), Git operations (creating branches, merging code), and execution of other skills such as tdd and code-review.
Sanitization: The instructions do not mention any sanitization or validation of the content retrieved from the issue tracker before it is used to direct the work of implementer subagents.

implement-spec