skills/udecode/plate/slate-ar-perf/Gen Agent Trust Hub

slate-ar-perf

Pass

Audited by Gen Agent Trust Hub on Jul 2, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to run various local project commands using pnpm (such as bench:targets:list and bench:targets:check) and node (specifically tooling/scripts/bench-targets.mjs). These operations are used to manage performance benchmarks and are restricted to the local development environment.
  • [DATA_EXFILTRATION]: Analysis of network operations shows only a reference to a local development server (http://localhost:3100). There is no evidence of data transmission to external or untrusted domains.
  • [PROMPT_INJECTION]: The skill incorporates external data from a local registry file (benchmarks/targets/slate-v2.json) and user-provided arguments. While this creates an indirect prompt injection surface, the risk is mitigated by the skill's focus on structured performance testing within a specific project context.
  • Ingestion points: Reads data from benchmarks/targets/slate-v2.json and user-provided $ARGUMENTS.
  • Boundary markers: None identified for distinguishing instruction from data.
  • Capability inventory: Execution of shell commands via pnpm and node scripts (referenced in SKILL.md).
  • Sanitization: Not explicitly described for external data interpolation.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 2, 2026, 05:10 PM
Security Audit — agent-trust-hub — slate-ar-perf