How iampattanayak/claude-humanizer works
Claude Humanizer is a prompt-as-code skill for revising AI-generated text in Claude Code. The documented workflow takes a draft through an initial cleanup, a self-critique audit, and a final revision; the audit targets a 24-pattern matrix spanning wording, structure, and typography. The README documents invocation and workflow, but no source excerpts are available, so the graph represents only those documented relationships and not internal implementation details.
Main entry points: SKILL.md Self-critique audit · README.md Anti-pattern matrix
Components 7
Humanizer workflow
- Skill protocol
SKILL.md - Draft cleanup
SKILL.md - Self-critique audit
SKILL.md - Anti-pattern matrix
README.md - Final revision
SKILL.md
Other components
- User
- Claude Code CLI
Connections 8
- User → Claude Code CLI: submits draft
- Claude Code CLI → Skill protocol: invokes
- Skill protocol → Draft cleanup: guides
- Draft cleanup → Self-critique audit: continues to
- Self-critique audit → Anti-pattern matrix: checks against
- Self-critique audit → Final revision: prompts revision
- Final revision → Claude Code CLI: returns result
- Claude Code CLI → User: presents result