swiftui-pro
Pass
Audited by Gen Agent Trust Hub on Oct 26, 2026
Risk Level: SAFEPROMPT_INJECTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill uses specific instructions to override the agent's internal training data by asserting the existence of speculative platform versions and hardware. It explicitly directs the agent not to treat these as speculative, which forces the model into a context that may generate non-functional code or report false issues based on non-existent APIs.
- Evidence: The Core Instructions section in
SKILL.mdstates: "iOS 27 exists, and is the default deployment target for new apps.", "Xcode 27.1 and the corresponding 27.1 platform releases are available. Do not dismiss their APIs as speculative.", and "iPhone Duo is a folding iPhone, with a small outer display and a large inner display." - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user-provided source code for review but does not provide instructions to mitigate risks from instructions embedded within that data, such as malicious comments designed to influence the agent's behavior.
- Ingestion points: The skill's primary function is to read and review Swift and SwiftUI code provided by the user (as stated in the
SKILL.mddescription). - Boundary markers: Absent. There are no instructions to the agent to treat user code content as data only or to ignore natural language instructions found within code comments.
- Capability inventory: The agent retains access to its full suite of tools (file operations, shell execution, etc.) while performing reviews under this skill's context.
- Sanitization: Absent. No filtering or escaping of content within the code being reviewed is defined.
Audit Metadata