a1-marketing-context

Pass

Audited by Gen Agent Trust Hub on Aug 11, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill operates under a strict execution framework called the "Context Spine," which ensures that all file modifications are authorized by explicit user intent and confined to designated repository-local paths.
  • [PROMPT_INJECTION]: The skill ingests data from external sources such as README.md and package.json to update repository context, creating a surface for potential indirect prompt injection.
  • Ingestion points: Extracts information from README.md, package.json, and legacy context files within the current repository.
  • Boundary markers: It automatically generates semantic markers ("How to Use This Context") to help downstream agents distinguish between protected product facts and overridable defaults.
  • Capability inventory: The skill has permissions to read repository documentation, write to the specific .agents/marketing-context.md path, and remove legacy context files during a user-confirmed migration.
  • Sanitization: It implements a "Confirmed Source Policy" that mandates the exclusion of tentative or speculative content, specifically filtering out text labeled as "hypotheses," "ideas," or "maybe."
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 11, 2026, 06:22 AM
Security Audit — agent-trust-hub — a1-marketing-context