agent-memory

Warn

Audited by Gen Agent Trust Hub on May 29, 2026

Risk Level: MEDIUMPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection through its memory retrieval mechanisms.\n
  • Ingestion points: Memory is retrieved using mem.recall() and mem.get_lessons() as described in SKILL.md.\n
  • Boundary markers: There are no delimiters or explicit instructions to treat retrieved memory as potentially untrusted data.\n
  • Capability inventory: The skill persists and recalls context which directly influences the agent's reasoning and future responses.\n
  • Sanitization: No sanitization or validation of the stored content is described.\n- [EXTERNAL_DOWNLOADS]: The skill recommends installation via clawdhub install agent-memory. This command downloads and installs code from an external registry.
Audit Metadata
Risk Level
MEDIUM
Analyzed
May 29, 2026, 04:28 PM
Security Audit — agent-trust-hub — agent-memory