The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes external, untrusted data from an issue tracker (issue bodies, comments, and PR diffs) to categorize and verify them. This data is interpolated into the agent's context and used to drive decision-making and command execution.
Ingestion points: Reads issue/PR bodies, comments, labels, and code diffs from the project's issue tracker (SKILL.md).
Boundary markers: No explicit delimiters or instructions to ignore embedded commands are specified for the external content being processed.
Capability inventory: The agent can post comments, modify labels, close issues, write to the filesystem (.out-of-scope/, GLOSSARY.md, ADRs), and execute shell commands/tests (SKILL.md, Step 3 and 5).
Sanitization: No sanitization or validation of the external content is mentioned before it is processed or used in verification steps.- [DYNAMIC_EXECUTION]: The skill explicitly instructs the agent to execute code from external sources during the verification phase. Specifically, it directs the agent to 'check out' a pull request and 'run the relevant tests or commands' to confirm the diff's claims. This constitutes a direct execution path for untrusted code provided by external contributors (SKILL.md, Step 3).- [COMMAND_EXECUTION]: The instructions require the agent to execute shell commands to interact with the repository, check out branches, and run test suites provided in external contributions (SKILL.md, Step 3).

triage