gundlach-bonds

Pass

Audited by Gen Agent Trust Hub on Aug 31, 2026

Risk Level: SAFENO_CODE
Full Analysis
  • [NO_CODE]: The skill consists entirely of Markdown files providing instructional templates for an AI agent. There are no scripts, binaries, or automated tasks that could execute on a host system.- [EXTERNAL_DOWNLOADS]: The skill includes links to well-known services and research sites (Spotify, DoubleLine, Lyn Alden). These are provided as informational resources for the user or agent to reference and do not involve automated downloads of executable content.- [INDIRECT_PROMPT_INJECTION]: While the skill instructs the agent to process external financial data and portfolio context, it lacks the tools or capabilities (such as network access or file system writes) to be exploited via data ingestion. It functions as a static judgment framework.
  • Ingestion points: User-provided inputs defined in SKILL.md (Asset, portfolio, theme, or market context; financial, market, regime, or structural data).
  • Boundary markers: Absent.
  • Capability inventory: None. No subprocess calls, network operations, or file-write capabilities are present in the skill.
  • Sanitization: Absent.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 31, 2026, 11:48 PM