park-this-for-later-with-enough-context

Pass

Audited by Gen Agent Trust Hub on Jul 22, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill is a purely instructional prompt template designed to help an agent summarize a conversation tangent. It contains no code, network requests, or sensitive file access.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided conversation data to generate its summary.\n
  • Ingestion points: Processes the current conversation and artifacts in context (SKILL.md).\n
  • Boundary markers: None explicitly defined to delimit external content from instructions.\n
  • Capability inventory: None; the skill explicitly disables model invocation (disable-model-invocation: true) and contains no subprocess calls, file operations, or network requests.\n
  • Sanitization: No sanitization or filtering of the ingested context is performed.\n
  • Risk Assessment: While the skill lacks boundaries and sanitization for ingested data, the total absence of tool-calling or system-level capabilities makes the risk of a malicious instruction being successfully executed negligible.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 22, 2026, 06:41 AM
Security Audit — agent-trust-hub — park-this-for-later-with-enough-context