skills/anfreire/wiki-spaces/ws-tend/Gen Agent Trust Hub

ws-tend

Pass

Audited by Gen Agent Trust Hub on Sep 10, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process content from markdown files within a user's wiki to perform audits, tag normalization, and cross-linking.
  • Ingestion points: Markdown files and metadata configuration files (e.g., _meta/taxonomy.md) located in the resolved wiki directory.
  • Boundary markers: The skill defines "owned" vs "external" trust scopes based on folder names (shared/), git submodules, and symlinks to prevent unauthorized modifications.
  • Capability inventory: Execution of the bundled scripts/ws.py script, shell commands for logging (printf), and file modification capabilities.
  • Sanitization: The skill enforces file size limits and structural validation but does not explicitly filter the natural language content for malicious instructions.
  • [COMMAND_EXECUTION]: The skill invokes a local Python script scripts/ws.py and standard shell utilities (printf, date) to perform its maintenance tasks. These executions are constrained to the local environment and the specific wiki directory resolved at runtime.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 10, 2026, 08:53 AM
Security Audit — agent-trust-hub — ws-tend