conversation-archaeologist

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to search and retrieve historical conversation data via conversation_search and recent_chats to analyze user style and context. This creates a surface where malicious instructions or untrusted data previously imported into past chats could affect the profile generation process.
  • Ingestion points: Historical chat data retrieved from conversation_search and recent_chats in SKILL.md.
  • Boundary markers: Absent. No clear delimiters or constraints are provided to ensure the agent treats old chat contents strictly as data rather than instructions.
  • Capability inventory: The skill produces an aggregated Markdown profile meant to be referenced by other skills.
  • Sanitization: No filtering or sanitization rules are outlined to handle potentially untrusted text inside historical conversations.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 01:01 PM
Security Audit — agent-trust-hub — conversation-archaeologist