The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted external data such as code, PR diffs, and architectural decisions.
Ingestion points: Data enters the agent's context through the main argument hint and file-reading operations in 'Code Review' and 'PR Review' modes.
Boundary markers: The skill instructions do not specify the use of delimiters (e.g., XML tags or triple backticks) or explicit 'ignore embedded instructions' warnings when the agent processes the subject under review.
Capability inventory: The skill has access to the Bash tool, which provides a significant capability surface if the agent were to be successfully influenced by malicious instructions embedded in the analyzed data.
Sanitization: There are no instructions for sanitizing, escaping, or validating the content of the files being reviewed before the agent processes them.

swing-review