agent-ready-av-campaign-skill
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted user-provided marketing content (drafts, emails, social posts) which could contain malicious instructions designed to influence the agent's behavior.
- Ingestion points: User-supplied campaign content provided during triggers such as "draft the pillar page" or "review this scenario page" (SKILL.md).
- Boundary markers: The skill uses XML-style delimiters (
<objective>,<quick_start>,<house_rules>,<core_content>) in its main instruction file to separate sections. - Capability inventory: The skill is instructed to write logs and analytics to the local filesystem at
~/.claude/skill-analytics/and~/.claude/skill-runs/. It also invokes external tools for brand checks and product catalog searches (SKILL.md, config.json). - Sanitization: The instructions do not specify any sanitization, escaping, or filtering of user-provided content before it is interpolated or evaluated against the campaign guidelines.
Audit Metadata