session-handoff

Pass

Audited by Gen Agent Trust Hub on Feb 19, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION] (LOW): Indirect Prompt Injection Surface. The skill is designed to ingest and process data from previous interactions, which creates a vulnerability if those interactions contained malicious instructions from untrusted sources (e.g., processed files or web content).
  • Ingestion points: Context is ingested through the <session-continuity> tag and the memory_sessions tool calls.
  • Boundary markers: The skill uses XML-style delimiters (<session-continuity>) to isolate session data, which helps the model distinguish context from instructions but is not a complete mitigation.
  • Capability inventory: The skill primarily uses memory retrieval tools (memory_sessions, memory_query). It does not contain scripts for arbitrary command execution or network requests.
  • Sanitization: The skill does not define specific sanitization routines for the data retrieved from memory or session files.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 19, 2026, 03:51 PM