Content
65%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
The content is highly actionable with executable commands and a clean flags reference, but it is slightly repetitive, lacks validation feedback loops in its batch workflows, and keeps everything inline with no progressive disclosure to separate files.
Suggestions
Add an explicit verify-before-push checkpoint to the parallel-worktree and autonomous-build workflows (e.g. run tests / `gh pr checks` before opening PRs).
Dedupe the --no-auto-update / --always-approve / pty+tmux guidance so it lives once (flags table + rules) rather than recurring across sections.
Offload the detailed PR-review patterns and full flags reference into a references/ file referenced one level deep from SKILL.md to improve progressive disclosure.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is mostly operational with no padded conceptual explanation, but key advice (e.g. --no-auto-update, --always-approve, pty+tmux) repeats across the flags table, deep dive, pitfalls, and rules sections and could be tightened. | 2 / 3 |
Actionability | Numerous copy-paste-ready `terminal(command="grok ...")` blocks, a complete flags table, config.toml, and concrete pitfall fixes provide fully executable guidance. | 3 / 3 |
Workflow Clarity | Sequences like the Read-Only Audit pattern have explicit validation checkpoints, but batch/destructive flows (parallel worktree fixes, autonomous --always-approve builds) jump straight to push/PR with no validate→fix→retry loop, capping the score at 2. | 2 / 3 |
Progressive Disclosure | Sections are well-organized, but with no bundle files the ~290-line SKILL.md keeps flags reference, PR-review, and worktree patterns inline rather than offloaded to one-level-deep references. | 2 / 3 |
Total | 9 / 12 Passed |