1k-new-skill

Pass

Audited by Gen Agent Trust Hub on Mar 4, 2026

Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill uses the WebFetch tool to retrieve documentation from code.claude.com and platform.claude.com. These are trusted domains associated with the AI service provider.
  • [PROMPT_INJECTION]: An indirect prompt injection surface is present as the skill is instructed to ingest external content and follow instructions or 'guidelines' extracted from it.
  • Ingestion points: WebFetch of documentation URLs defined in the 'Fetch latest best practices' section of SKILL.md.
  • Boundary markers: Absent; no delimiters or 'ignore' warnings are provided when processing external content.
  • Capability inventory: The skill has access to Read, Grep, Glob, WebFetch, and Write tools, as well as the ability to execute shell commands.
  • Sanitization: Absent; content from the external URLs is used directly to guide the generation of new skills.
  • [COMMAND_EXECUTION]: The skill instructs the agent to execute several shell commands for file management and Git workflows (ls, mkdir, mv, git add/commit). It also invokes a local Python script located at 'development/skills-analysis/analyze-skills-tokens.py' to perform token count verification on newly created skills. These operations are consistent with the skill's purpose as a development tool.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 4, 2026, 04:17 AM