skills/anomalyco/opencode/effect/Gen Agent Trust Hub

effect

Pass

Audited by Gen Agent Trust Hub on Sep 17, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructs the agent to clone the effect-smol repository from the Effect-TS organization on GitHub if the reference directory is missing. This repository provides the source of truth for API usage and naming patterns for the Effect v4 library.
  • [INDIRECT_PROMPT_INJECTION]: The skill directs the agent to analyze the contents of the cloned repository to inform its responses and code generation, which introduces a potential surface for indirect prompt injection if the external source were compromised.
  • Ingestion points: Documentation, examples, and source files within the .opencode/references/effect-smol/ directory.
  • Boundary markers: Absent; the skill does not specify delimiters or instructions to ignore embedded commands within the reference material.
  • Capability inventory: Includes file system read access, repository search, and git clone capabilities.
  • Sanitization: Absent; content from the repository is treated as an authoritative reference without explicit validation or filtering.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 17, 2026, 12:23 AM
Security Audit — agent-trust-hub — effect