skills/xsavikx/okf-skills/okf-graphql/Gen Agent Trust Hub

okf-graphql

Pass

Audited by Gen Agent Trust Hub on Sep 12, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection as it ingests and processes user-provided GraphQL SDL (Schema Definition Language) documents.
  • Ingestion points: The loadSchema function in graphql.go reads the content of files provided through the --schema CLI flag.
  • Boundary markers: The skill converts GraphQL types and operations into Markdown files without implementing specific boundary delimiters that would prevent an agent from interpreting comments or descriptions within the schema as instructions.
  • Capability inventory: Across main.go and graphql.go, the skill is capable of directory creation, file reading, and file writing to the local system to generate the OKF bundle.
  • Sanitization: The tool utilizes okf.SanitizeCell and a sanitizeLine function to clean strings and prevent table breakage in the resulting Markdown concepts.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 12, 2026, 08:04 AM
Security Audit — agent-trust-hub — okf-graphql