slow-novel-storyteller

Pass

Audited by Gen Agent Trust Hub on Oct 19, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to process external author materials and research, creating a theoretical attack surface for indirect prompt injection if those sources contain malicious instructions.
  • Ingestion points: The workflow ingests author drafts, notes, beatboards, research material, and legacy AI drafts as described in SKILL.md and source-authority.md.
  • Boundary markers: No specific boundary markers or "ignore embedded instructions" delimiters are defined for these inputs.
  • Capability inventory: The workflow includes writing to local files (PDF, TeX, logs, and backups), performing database queries for state management, and executing atomic synchronization for shared folders (SKILL.md, session-architecture.md).
  • Sanitization: No specific sanitization or validation protocols for the ingested literary content are mentioned.
  • [DATA_EXPOSURE_AND_EXFILTRATION]: The skill documentation includes security-positive instructions, specifically advising the agent to keep prompts, databases, logs, and backups private and to verify file integrity using checksums during synchronization.
  • [SAFE]: No malicious patterns related to remote code execution, command injection, obfuscation, or privilege escalation were identified. The skill focuses on workflow orchestration and literary guidance.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 19, 2026, 06:27 PM