authoring-scouts

Pass

Audited by Gen Agent Trust Hub on Jul 28, 2026

Risk Level: SAFE
Full Analysis
  • [PROMPT_INJECTION]: The skill instructs agents on authoring and analyzing other AI agent skills, creating an indirect prompt injection surface. 1. Ingestion points: Reading skill bodies and references via posthog:skill-get. 2. Boundary markers: Direct instructions to treat ingested text as evidence, not commands, and to ignore steering attempts. 3. Capability inventory: Capabilities include skill creation/modification and inbox report authoring. 4. Sanitization: Clear guidance on redacting personal information from report excerpts.
  • [COMMAND_EXECUTION]: The documentation references standard PostHog CLI and Django management commands (hogli, sync_signals_scout_skills) used for developer workflows in an instructional context.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 28, 2026, 01:34 PM
Security Audit — agent-trust-hub — authoring-scouts