Content
87%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
The body is a concise, highly actionable multi-stage verification workflow with executable commands and a useful report template. Its main weakness is uneven validation feedback loops across the later stages.
Suggestions
Add explicit validate->fix->retry checkpoints to the lint, security, and diff-review stages so every stage has a clear pass/fail decision gate.
Clarify the coverage threshold enforcement (e.g. 'fail the stage if coverage < 80%') so the test stage has a concrete go/no-go rule.
Specify what 'critical' type errors vs warnings mean for the type-check stage's stop/continue decision.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean and command-focused, with no padding or explanation of concepts Claude already knows; every section earns its place with concrete shell commands, matching the lean-and-efficient anchor. | 3 / 3 |
Actionability | Each stage provides concrete, executable commands with language-specific variants (npm/pnpm, tsc/pyright, ruff) and a copy-paste-ready report template, matching the fully-executable anchor. | 3 / 3 |
Workflow Clarity | A clear 6-stage sequence with an explicit stop-and-fix checkpoint after the build stage exists, but later stages (lint, security scan, diff review) lack explicit validate->fix->retry feedback loops, matching the 'sequence present but checkpoints missing or implicit' level 2 rather than the consistent-checkpoint level 3. | 2 / 3 |
Progressive Disclosure | Content is well-organized into clearly labeled sections (six numbered stages, output format, continuous mode, hooks integration) within a single appropriately-scoped file with no deep or nested references, matching the clear-overview anchor. | 3 / 3 |
Total | 11 / 12 Passed |