Content
86%Weight 40%Scale 1-5Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
The content is concise, well-structured, and keeps the RFC template as the external source of truth with a clear workflow and validation checklist. The weakest areas are minor: a conditional lint step without an explicit fix-retry loop and a few under-specified edge cases.
Suggestions
Add an explicit validate-then-fix-retry loop for the Markdown lint check (e.g., "if lint fails, fix formatting and re-run until it passes") rather than leaving it conditional.
Specify a concrete fallback when the RFC author is unknown beyond "leave the template placeholder" (e.g., default value or a prompt to ask the user).
Clarify how to detect "the next available NNNN" (e.g., list rfc/NNNN-* dirs and increment the highest number) so the step is fully executable.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean and efficient, organized into Workflow, Writing Standards, and Validation sections with no padding or explanation of concepts Claude already knows; every line adds task-specific guidance. | 5 / 5 |
Actionability | Concrete guidance with specific paths and placeholders ("Create rfc/NNNN-short-title/README.md", "state: draft") is actionable, but a few items like the conditional lint check and "author unknown" handling leave minor gaps. | 4 / 5 |
Workflow Clarity | A clear numbered sequence (steps 1-8) plus a Validation checklist and an ask-user feedback loop are present, but the lint check is conditional with no explicit fix-retry loop, leaving a minor validation gap. | 4 / 5 |
Progressive Disclosure | The under-50-line body is well-organized into clear sections and keeps the template and process docs as one-level-deep external sources of truth (rfc/README.md, rfc/0000-template/README.md), giving easy navigation with no nested references. | 5 / 5 |
Total | 18 / 20 Passed |