event-store-semantics

Pass

Audited by Gen Agent Trust Hub on Aug 19, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill provides documentation and behavioral rules for an event store implementation. It does not contain executable code, scripts, or instructions that perform network operations or access sensitive system files.\n- [INDIRECT_PROMPT_INJECTION]: The skill describes an attack surface involving the ingestion of third-party Nostr events. While the documentation itself is benign, an agent utilizing these rules to process untrusted data should implement appropriate sanitization and boundary markers.\n
  • Ingestion points: Data enters through the IEventStore interface and SQLite reference implementation described in SKILL.md.\n
  • Boundary markers: No specific prompt boundary markers or 'ignore' instructions are defined in this behavioral spec.\n
  • Capability inventory: The skill outlines database query and write operations (SQL), but does not itself execute shell commands or system-level processes.\n
  • Sanitization: The specification mentions NIP-50 search extension stripping (STORE-S01) and FTS5 grammar enforcement (STORE-S04), providing some degree of input normalization.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 19, 2026, 07:48 PM
Security Audit — agent-trust-hub — event-store-semantics