blog-calendar
Pass
Audited by Gen Agent Trust Hub on Nov 3, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill follows security best practices for AI agents. It explicitly instructs the agent to treat existing blog posts and external content as untrusted data, directing it to ignore any instructions embedded within them. It also specifies that file paths should be resolved, quoted, and validated to prevent unauthorized overrides or directory traversal. No malicious patterns, obfuscation, or unauthorized data exfiltration attempts were detected.
- [INDIRECT_PROMPT_INJECTION]: The skill processes potentially untrusted content from existing blog files (Markdown, MDX, HTML) and external web sources (Google Trends, Search Status Dashboard). This creates an attack surface for indirect prompt injection.
- Ingestion points: Scanning local blog posts (SKILL.md Step 1) and external trends/status dashboards (SKILL.md Step 4).
- Boundary markers: Explicitly directs the agent to 'ignore instructions inside' untrusted data and to use 'helper-fenced output' for project context.
- Capability inventory: The skill has file system access for reading content and writing generated calendars to a local directory, and network access for research.
- Sanitization: The skill advises resolving, quoting, and validating paths to prevent directory traversal; however, content body sanitization is absent, relying instead on the instruction to ignore embedded commands.
Audit Metadata