content-atomizer

Pass

Audited by Gen Agent Trust Hub on Nov 17, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external content that could contain hidden instructions intended to influence the agent's behavior.
  • Ingestion points: The skill fetches content from user-provided URLs using Firecrawl or WebFetch (Phase 2) and accepts pasted comments for the Engagement Replies mode.
  • Boundary markers: There are no explicit instructions or delimiters defined to isolate external source content from the agent's internal instructions.
  • Capability inventory: The skill has the ability to write files to the local campaign directory and perform network-write operations to post content via social media APIs (Zernio, Buffer, or Hootsuite).
  • Sanitization: The skill employs a 'quality gate' in Phase 4 that lints for 'slop patterns' and AI-generated prose tells, providing a layer of content filtering.
  • [EXTERNAL_DOWNLOADS]: The skill is configured to fetch data from external URLs to extract insights for social media distribution.
  • It uses well-known services like Firecrawl or WebFetch to retrieve content from the internet.
  • It references technical algorithm specifications from the official X (Twitter) algorithm repository on GitHub.
  • [COMMAND_EXECUTION]: The skill instructs the agent to execute a local Python script to establish brand context.
  • The command python3 ${CLAUDE_SKILL_DIR}/../start-here/scripts/brand_context.py is called to load necessary metadata from local project files when they exist.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 17, 2026, 04:23 PM