checking-skill-portability

Warn

Audited by Gen Agent Trust Hub on Nov 19, 2026

Risk Level: MEDIUMREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The skill provides instructions to download and execute code from an external, third-party GitHub repository using the uvx tool runner.
  • Evidence: In SKILL.md, the command uvx --from "git+https://github.com/agentskills/agentskills@f130f348f502d9804278a617f86929846896d2e9#subdirectory=skills-ref" skills-ref validate <skill-directory> installs and runs a validator script directly from a remote source.
  • Although the command is pinned to a specific commit hash (f130f348f502d9804278a617f86929846896d2e9), the source repository belongs to an unverified third party.
  • [EXTERNAL_DOWNLOADS]: The skill's operation depends on fetching external components from the internet.
  • Evidence: The frontmatter compatibility field in SKILL.md explicitly states that the validator tool "needs network access the first time" it is run.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data (other skill packages), which could contain malicious instructions designed to manipulate the agent or the validation tool.
  • Ingestion points: The skill reads all content within the target <skill-directory> mentioned in SKILL.md.
  • Boundary markers: Absent. No specific delimiters or instructions to ignore embedded commands are used when analyzing the target files.
  • Capability inventory: The skill has the capability to execute shell commands via uvx as part of its primary workflow.
  • Sanitization: Absent. The skill does not describe any validation or sanitization of the input files before they are processed by the agent or the shell-based validator.
  • [COMMAND_EXECUTION]: The skill utilizes a shell command block to execute its main validation logic.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Nov 19, 2026, 01:21 AM