swift-codable

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFEDATA_EXFILTRATIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [DATA_EXFILTRATION]: The file SKILL.md contains Swift code examples that utilize URLSession to perform network operations, such as fetching data from https://api.example.com/users/.
  • [EXTERNAL_DOWNLOADS]: The References section in SKILL.md links to external documentation hosted on sosumi.ai, which is a third-party domain not included in the trusted vendors list.
  • [INDIRECT_PROMPT_INJECTION]: The skill provides patterns for ingesting and parsing external data, which presents a potential surface for indirect prompt injection if the parsed data is subsequently used in LLM prompts.
  • Ingestion points: The fetchUser function in SKILL.md demonstrates fetching data from an external API URL.
  • Boundary markers: The implementation uses JSONDecoder to enforce structure on the incoming data, which acts as a boundary against unstructured content.
  • Capability inventory: The skill documents the use of URLSession for network access and JSONDecoder for data processing.
  • Sanitization: Type-safe decoding into Swift structs provides validation of the incoming data schema.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 12:59 AM
Security Audit — agent-trust-hub — swift-codable