The Agent Skills Directory

[SAFE]: The skill ensures input safety by using 'git rev-parse' to validate user-supplied commit or branch references before they are used in comparison commands, mitigating potential command injection.
[SAFE]: The architecture employs parallel sub-agents for different review axes, which isolates the analysis context and prevents instructions in one part of the code from interfering with other review tasks.
[PROMPT_INJECTION]: The skill has an indirect prompt injection surface as it processes untrusted diffs and spec files. (1) Ingestion points: Data enters via 'git diff' output, 'git log', and documentation files. (2) Boundary markers: No specific delimiters or safety instructions are defined for the sub-agent prompts. (3) Capability inventory: Capabilities are limited to 'git' read-only commands and coordinated sub-agent prompts. (4) Sanitization: The skill does not explicitly sanitize or escape the content before it is processed by the AI sub-agents.

code-review