loop-research

Pass

Audited by Gen Agent Trust Hub on Sep 5, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion and analysis of untrusted data from open-source repositories and research papers for benchmarking purposes.\n
  • Ingestion points: preset.md defines a cohort collection process for harvesting source implementations and documentation from external technical sources.\n
  • Boundary markers: No explicit delimiters or safety instructions are defined in the authoring prompts to isolate external data from agent context.\n
  • Capability inventory: The skill authors code for the loop-graph executor, which is capable of running shell commands and evaluation harnesses, though this skill itself does not invoke those tools.\n
  • Sanitization: There are no explicit instructions for sanitizing or escaping the content retrieved from external sources before it is processed by the agent.\n- [DYNAMIC_EXECUTION]: The skill generates executable templates and operation files intended for a downstream runtime environment.\n
  • Evidence: SKILL.md and README.md describe the compilation of executor, supervisor, and operations files using templates found in the ../loop-graph/templates/ directory.\n
  • Constraint: The skill contains an explicit safety guard: "Do not execute or resume generated runtime node files from this authoring skill," which serves to maintain a clear boundary between the authoring process and runtime execution.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 5, 2026, 02:38 PM
Security Audit — agent-trust-hub — loop-research