yzr-llm-wiki-management

Pass

Audited by Gen Agent Trust Hub on Aug 17, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill includes several Python utilities (lint_wiki.py, ingest_diff.py, check_wiki_fixtures.py) that perform local Git operations and structural verifications using the subprocess module. These commands are executed with structured argument lists and are focused on retrieving repository status and metadata for wiki maintenance.
  • [EXTERNAL_DOWNLOADS]: The skill documentation provides instructions for the agent to assist users in cloning external Git repositories into a specific raw/external/ directory to serve as data sources. This is a core feature for technical documentation wikis and is designed with clear role boundaries where the LLM treats these sources as read-only during maintenance tasks.
  • [SAFE]: The skill implements proactive security measures, notably a check in the linter that forbids the use of absolute file paths (e.g., /etc/passwd) in document metadata. This prevents the accidental ingestion or exposure of sensitive files outside the designated wiki root. Additionally, the skill's architecture strictly separates user-controlled raw data from LLM-generated summaries to maintain data integrity.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 17, 2026, 09:42 AM
Security Audit — agent-trust-hub — yzr-llm-wiki-management