The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill relies on GitHub issues as the source of tasks and specifications. These issues are external, untrusted data sources that can be manipulated by anyone with repository access to include malicious instructions.\n
Ingestion points: Data is ingested through gh issue view commands which retrieve issue metadata and content (Step 1).\n
Boundary markers (absent): The instructions do not define clear delimiters or use prompt engineering techniques to distinguish between the agent's instructions and the content of the issues.\n
Capability inventory: The skill allows for the creation of branches, merges, file writes outside the repository, and the invocation of subagents for implementation and testing tasks (Steps 2, 3, 4, 5, 8).\n
Sanitization (absent): There is no mention of filtering, escaping, or validating the issue content before it is processed by the agent or its subagents.\n- [COMMAND_EXECUTION]: The skill uses the gh CLI tool to interact with GitHub services and manages local repository state via git worktrees and branches. It also calls platform-provided tools for 'tdd' and 'code-review'.

implement-spec