kotlin-tooling-kotlin-toolchain-plugin-authoring

Pass

Audited by Gen Agent Trust Hub on Sep 11, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill documents an architecture where local plugins ingest external data to drive build-time tasks, which is a potential surface for indirect instructions to influence agent actions.
  • Ingestion points: Plugins ingest data through the @Configurable settings interface (populated from module.yaml) and via System.getenv() calls within task actions (SKILL.md, references/examples.md).
  • Boundary markers: The code snippets provided do not use explicit delimiters or "ignore instructions" markers when processing these inputs.
  • Capability inventory: Task actions documented in the skill have the capability to write to the filesystem (writeText), delete directories (deleteRecursively), and interact with the local git repository (GitRepo) (references/examples.md).
  • Sanitization: The documentation explicitly advises that users should "validate a value before using it in a file path or process argument," but the accompanying code examples do not demonstrate sanitization or validation of the external inputs.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 11, 2026, 02:41 PM
Security Audit — agent-trust-hub — kotlin-tooling-kotlin-toolchain-plugin-authoring