token-budget-advisor

Pass

Audited by Gen Agent Trust Hub on Jun 14, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill consists of instructional text and reference documentation. No executable scripts, subprocess calls, or system modifications were found across any of the files.
  • [INDIRECT_PROMPT_INJECTION]: The skill instructions specifically guide the agent to ingest and analyze untrusted external content such as multiple uploaded documents and web research findings. This represents a known attack surface.
  • Ingestion points: The framework explicitly mentions handling 'Multiple file uploads (3+ documents)', 'Web search result', and 'Web fetch (full article)' in SKILL.md.
  • Boundary markers: The skill does not provide specific instructions to use delimiters or 'ignore' tags when processing these inputs to mitigate potential malicious instructions embedded in documents.
  • Capability inventory: The skill is intended to be used in workflows that involve reading local files, conducting web searches, and generating long-form artifacts (reports, presentations).
  • Sanitization: No sanitization methods or input validation strategies are suggested for the external content before processing.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 14, 2026, 11:11 AM
Security Audit — agent-trust-hub — token-budget-advisor