The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted code diffs, which creates a potential ingestion surface for malicious instructions hidden within comments or strings of the code being reviewed.
Ingestion points: The skill reviews external code diffs provided by the user during a review workflow.
Boundary markers: The instructions do not currently specify the use of delimiters or specific warnings to ignore instructions embedded in the code diffs.
Capability inventory: The skill configuration explicitly includes disable-model-invocation: true, which prevents the agent from calling external tools, performing network requests, or writing to the file system, effectively neutralizing the impact of any injected instructions.
Sanitization: There is no explicit sanitization of the input code mentioned in the instructions.
[SAFE]: The skill's primary function is textual analysis. It does not include any executable scripts, remote downloads, or credential requirements. The references to external design tools and documentation (e.g., animations.dev, easing.dev) are plain-text informational links to well-known community resources.

review-animations