The Agent Skills Directory

[NO_CODE]: The skill consists exclusively of markdown instructions for an AI agent to perform a persona-based task. It does not contain any scripts, binary files, or configuration that would execute shell commands or interact with the system at the operating system level.
[INDIRECT_PROMPT_INJECTION]: The skill's primary function is to analyze external, potentially untrusted code (diffs and related files). This presents a surface for indirect prompt injection where instructions could be hidden in the code being reviewed. However, the risk is mitigated as the skill does not utilize tools with side effects like network access, file system modification, or process execution. Ingestion points: Code changes (diffs) and imported source files (SKILL.md). Boundary markers: The agent is instructed to first summarize the change's intent and read context files, which helps maintain focus on the review objective. Capability inventory: There are no subprocess calls, file-write operations, or network requests present in the skill definition. Sanitization: The skill does not implement specific sanitization but operates within the standard safety boundaries of the AI agent.

code-review-commons