The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to perform several shell-based git operations, including git fetch, git checkout, git remote add, and git diff. These commands are used to manage local branches and compare changes against an upstream target.
[EXTERNAL_DOWNLOADS]: The workflow involves fetching code from remote repositories (e.g., gitcode.com, github.com) using git fetch. While the instructions provide examples targeting the author's own repositories (Ascend/RecSDK), the logic is designed to pull from arbitrary URLs provided in pull request links.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external PRs, including code diffs and commit metadata. This creates an attack surface where a malicious PR could attempt to influence the agent's behavior during the review process.
Ingestion points: Pull request diffs and source code fetched from remote branches (defined in SKILL.md).
Boundary markers: Absent; there are no specific instructions or delimiters provided to help the agent distinguish between code-to-be-reviewed and potential embedded instructions.
Capability inventory: The skill has access to shell command execution (git) and full file system reading capabilities within the repository context.
Sanitization: None; the agent processes raw output from git diff without filtering or escaping content.

refactor-code-review