The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data in the form of pull request diffs and source code, creating a surface for indirect prompt injection attacks where malicious instructions could be embedded in the code being reviewed.
Ingestion points: The skill accepts diffs and code snippets for review as part of the core functionality described in SKILL.md and README.md.
Boundary markers: The instructions do not define specific delimiters or instructions to ignore embedded commands within the diffs being processed.
Capability inventory: The skill's capabilities are limited to generating text output for PR comments. It lacks the ability to execute subprocesses, perform network operations, or write to the filesystem.
Sanitization: No input sanitization or validation logic is present in the instructions.

caveman-review