workflow-from-chats

Pass

Audited by Gen Agent Trust Hub on Jun 21, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it processes untrusted chat transcripts to generate persistent artifacts like rules and skills.\n
  • Ingestion points: Reads Cursor chat transcripts and subagent logs (SKILL.md).\n
  • Boundary markers: Absent; the workflow does not use specific delimiters or instructions to prevent the agent from obeying commands embedded within the transcripts being analyzed.\n
  • Capability inventory: The skill identifies the ability to write and edit files (new skills, rules, workflow docs) based on extracted content.\n
  • Sanitization: Absent; while it filters for 'reusable guidance,' it does not perform security validation or sanitization of the extracted content before writing it to the filesystem.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 21, 2026, 05:33 PM
Security Audit — agent-trust-hub — workflow-from-chats