The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external sources, such as issue tracker bodies and comments, to drive its automated implementation and review workflow. This content is passed to subagents for code generation and execution, creating a vulnerability surface where malicious instructions embedded in a ticket could influence the generated code or agent behavior.
Ingestion points: Section 2 describes resolving work items by fetching the full body, comments, and acceptance criteria from a tracker (e.g., Jira).
Boundary markers: The skill lacks explicit delimiters or instructions to the subagents to treat ticket content strictly as data rather than instructions.
Capability inventory: Subagents orchestrated by this skill have the ability to modify local files, commit to git, push to remotes, open pull requests, and modify tracker states (Sections 3, 4, and 5).
Sanitization: No sanitization, filtering, or validation of the fetched ticket content is performed before it is used to guide implementation.
[EXTERNAL_DOWNLOADS]: The skill instructs users to install external workflow dependencies from third-party sources, specifically recommending mattpocock/skills via npx and Ponytail skills from a personal GitHub repository.
[COMMAND_EXECUTION]: The skill performs automated git operations, including branch management, committing changes, pushing to remote repositories, and interacting with repository host APIs to create draft pull requests.

ship-ticket