skd-compile
Pass
Audited by Gen Agent Trust Hub on Nov 2, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill implements a data ingestion surface through its JSON DSL, which supports a shorthand (
@file) for reading external files and embedding their content into the generated XML. - Ingestion points: Input parameters
-DefinitionFileand-Valueinskd-compile.ps1andskd-compile.py, as well as referenced file paths within the JSON content. - Boundary markers: Absent; the skill does not use delimiters to isolate external content or warn the agent to ignore instructions embedded in data.
- Capability inventory: The skill performs file-read operations (
Get-Content,ReadAllBytes) and file-write operations (WriteAllText) to generate the schema file. - Sanitization: Absent; the logic resolving
@paths does not restrict which files can be read, creating a potential surface for reading sensitive local files if the agent is instructed to process an untrusted JSON definition.
Audit Metadata