skills/openclaw/skills/serp-analysis/Gen Agent Trust Hub

serp-analysis

Pass

Audited by Gen Agent Trust Hub on Feb 23, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill is composed of markdown instructions and metadata with no executable scripts or dangerous commands detected. All external tool references are abstract placeholders rather than direct code.
  • [PROMPT_INJECTION]: The skill defines a surface for indirect prompt injection as it processes external search engine data.
  • Ingestion points: Search result snippets, titles, and ranking page content (SKILL.md).
  • Boundary markers: The instructions use structured markdown templates, tables, and headers to compartmentalize data.
  • Capability inventory: There are no subprocess calls, system command executions, or network operations defined in the skill logic.
  • Sanitization: The workflow relies on natural language boundary separation and optional manual data entry to mitigate potential instruction overrides from search data.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 23, 2026, 08:46 AM