Content
57%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 body is strong on concrete, executable commands — most sections are copy-paste bash a user could run directly. Its weaknesses are structural: heavy duplication across sections, monolithic inlining of material (integrations, automation, prompt patterns) that belongs in separate reference files, and automated YOLO workflows that proceed with no validation checkpoints.
Suggestions
Split the Vim/VS Code/tmux integrations, automation scripts, and prompt-engineering patterns into separate reference files (e.g., references/integrations.md, references/automation.md) and keep SKILL.md as a lean overview with one-level-deep pointers.
Add validation/verification steps to the automated --yolo workflows (e.g., review generated files, run tests, checkpoint before executing batch changes) before allowing autonomous execution to proceed.
Deduplicate the repeated /compress, /checkpoint, and YOLO-toggle guidance into a single section and remove unverifiable commands like "gemini context size" and the shell-loop "gemini /compress".
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is mostly executable commands rather than padded prose, but is noticeably redundant: /compress, /checkpoint, YOLO toggling, and session-save logic each appear in multiple sections, and the prompt-engineering, automation, and integration catalogs inflate the file well past what the core workflow needs. This matches anchor 3 ("could be tightened") better than anchor 4, since the duplication goes beyond minor trimming. | 3 / 5 |
Actionability | Nearly every section is copy-paste-ready bash with concrete flags and slash commands (gemini -i "...", --checkpointing, @./src/... file references, /compress, tmux recipes). Minor gaps keep it below anchor 5: some dubious invocations (gemini context size, gemini /compress invoked from a shell loop, gemini checkpoint save) and the Vim mapping ":w !gemini -p \"... $(cat)\"" will not expand as written. | 4 / 5 |
Workflow Clarity | Sessions are presented as runnable snippets but as a catalog rather than sequenced workflows with checkpoints, and the automated --yolo workflows (full autonomous development cycles, batch file generation) include no validation or verification steps. Per the rubric, batch/automated operations without validation cap workflow clarity at 3. | 3 / 5 |
Progressive Disclosure | There are no bundle files: the entire catalog (interactive commands, prompt patterns, Vim/VS Code/tmux integrations, automation scripts) is inlined in a ~530-line SKILL.md with no references to separate files. Real section headers exist and navigation is possible, matching anchor 3 ("structure present; content that should be separate is inline") rather than anchor 2, whose example lacks any section structure. | 3 / 5 |
Total | 13 / 20 Passed |