client-deliverables

Pass

Audited by Gen Agent Trust Hub on Nov 12, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied evidence and analysis to generate consulting reports and presentations. This establishes an ingestion surface for untrusted data that could potentially influence agent behavior if the source content contains embedded instructions.
  • Ingestion points: User-provided context, evidence base, and analysis mentioned in SKILL.md.
  • Boundary markers: The instructions lack explicit delimiters or markers to isolate user data from the agent's primary instructions.
  • Capability inventory: The skill utilizes tool-specific implementations for document generation and executes local JavaScript for chart calculations.
  • Sanitization: The skill includes robust verification steps such as reconciling figures across artifacts, checking title-exhibit agreement, and inspecting rendered pages for accuracy.
  • [DYNAMIC_EXECUTION]: The skill includes a local JavaScript module (scripts/consulting-charts.cjs) to perform geometric calculations for waterfall charts. The agent is directed to use this script to generate editable presentation shapes, which is a core function of the skill.
  • [COMMAND_EXECUTION]: The skill provides implementation examples for programmatically generating PowerPoint files using the PptxGenJS library. This intended functionality allows the agent to transform analysis into finished artifacts.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 12, 2026, 12:44 PM