The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes external inputs in the form of specifications and issue tracker tickets to drive a complex automated workflow.
Ingestion points: The skill reads a "spec" and "tickets" from an issue tracker or user-provided context as described in SKILL.md (Step 1).
Boundary markers: There are no explicit instructions to use delimiters or "ignore embedded instructions" warnings when passing these inputs to the exploration, implementer, or merger subagents.
Capability inventory: The workflow includes repository management (git branches, worktrees, merges) and the invocation of the tdd and code-review tools.
Sanitization: The instructions do not mention any sanitization, validation, or filtering of the ticket content before it is used to influence the agent's actions.
[COMMAND_EXECUTION]: The skill automates several filesystem and repository operations based on the external task graph.
Evidence: The skill creates branches, manages git worktrees, performs merges, and executes the tdd tool (Step 4, Step 5, Step 7). While these are intended development tasks, they are executed autonomously based on potentially untrusted external input.

implement-spec