mem-promote

Pass

Audited by Gen Agent Trust Hub on Jun 15, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill exhibits a vulnerability to indirect prompt injection by ingesting data from daily log files that may contain untrusted content.
  • Ingestion points: The skill scans .memories/YYYY-MM-DD.md (project-level) or memories/YYYY-MM-DD.md (user-level) for entries marked as [Candidate].
  • Boundary markers: There are no specific instructions or delimiters defined to ensure the agent ignores potentially malicious instructions embedded within the source notes being processed.
  • Capability inventory: The skill possesses the capability to write to ~/.agents/AGENTS.md (global durable memory) and project files (AGENTS.md or CLAUDE.md), which allows injected instructions to achieve persistence across different sessions and projects.
  • Sanitization: The process relies on a mandatory human-in-the-loop confirmation of a diff/plan before writing to durable files, which serves as a primary defense but is not an automated sanitization or filtering mechanism for the content itself.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 15, 2026, 06:42 AM
Security Audit — agent-trust-hub — mem-promote