kelly-legal-precedent-desk

Pass

Audited by Gen Agent Trust Hub on Oct 9, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes legal research data from external sources and local payloads.
  • Ingestion points: scripts/create_research_batch.mjs reads JSON research payloads; the AirApp reads from Busabase Bases (items, entities, checks).
  • Capability Inventory: scripts/export_research_pack.mjs writes Markdown, JSON, and CSV files to the local exports/ directory. scripts/create_research_batch.mjs and scripts/execute_decisions.mjs perform writes to the Busabase platform.
  • Boundary Markers: The skill enforces an approval-gated review queue, requiring a human lawyer to approve research packs before they are considered valid for matter strategy or export.
  • Sanitization: The frontend application utilizes escapeHtml within app/js/review-views.js to sanitize content rendered in the browser UI.
  • [EXTERNAL_DOWNLOADS]: The skill utilizes standard Node.js dependencies.
  • The skill depends on busabase-sdk@0.80.0, hono@4.12.29, and @hono/node-server@2.0.8 for its core functionality and local gateway operations.
  • [COMMAND_EXECUTION]: The skill includes trusted scripts for administrative tasks.
  • Scripts like scripts/setup.mjs and scripts/publish_airapp.mjs are used for provisioning and deploying the skill's resources to the Busabase platform using environment-stored credentials.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 9, 2026, 01:37 PM