Skills
skills/vvkmnn/claude-gladiator-mcp/claude-gladiator/Gen Agent Trust Hub

claude-gladiator

Pass

Audited by Gen Agent Trust Hub on Apr 1, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill demonstrates a surface for indirect prompt injection because it captures context from tool outputs and user input and later incorporates it into agent configuration files (rules and hooks).\n
  • Ingestion points: Data enters through the gladiator_observe tool via the summary and context fields in SKILL.md.\n
  • Boundary markers: No explicit delimiters or instructions are used to separate captured external data from future system-level instructions.\n
  • Capability inventory: The gladiator_reflect tool has the capability to scan and suggest updates to configuration files in ~/.claude/rules/, ~/.claude/hooks/, and ~/.claude/skills/.\n
  • Sanitization: The instructions do not specify any validation or sanitization for the captured error messages or user corrections before they are stored.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 1, 2026, 12:57 PM