signup-flow-cro

Pass

Audited by Gen Agent Trust Hub on Apr 29, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No malicious patterns, obfuscation, or unauthorized data access behaviors were detected. The skill instructions are purely diagnostic and advisory.
  • [PROMPT_INJECTION]: The skill contains an indirect prompt injection surface by instructing the agent to read local project configuration files.
  • Ingestion points: Reads .agents/product-marketing-context.md and .claude/product-marketing-context.md in SKILL.md.
  • Boundary markers: Absent; no specific delimiters or warnings are provided for the ingested content.
  • Capability inventory: No capabilities for command execution, network access, or file modification are defined in the skill.
  • Sanitization: Absent; the ingested content is not validated or sanitized.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 29, 2026, 11:09 PM
Security Audit — agent-trust-hub — signup-flow-cro