Content
77%Weight 40%Scale 1-3Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
A well-structured, highly actionable cleanup workflow with strong validation and error handling for destructive operations. Its main weakness is monolithic inline content with no progressive disclosure into bundle reference files.
Suggestions
Move the per-IaC-tool destroy-command table and recommended /loop intervals into a references/ file (e.g. destroy-commands.md), keeping SKILL.md an overview with a clearly signaled one-level-deep link.
Consolidate the repeated "NEVER auto-destroy" safety directive into a single prominent callout instead of restating it in Steps 3, 4, and /loop Support.
De-duplicate the /loop usage guidance between Step 7's summary and the dedicated /loop Support section.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Operational specifics are mostly tight, but the "NEVER auto-destroy" safety directive is restated four times and /loop guidance is repeated across Step 7 and its own section, so it could be trimmed. | 2 / 3 |
Actionability | Provides executable, copy-paste-ready commands (e.g. `gh issue list --label "arn-infra-cleanup"`, `tofu destroy -target=...`) and a concrete per-IaC-tool destroy-command table with real syntax. | 3 / 3 |
Workflow Clarity | A clearly sequenced 7-step workflow with explicit confirmation checkpoints before destructive ops and thorough error-recovery feedback loops (state locks, partial failures, already-destroyed resources). | 3 / 3 |
Progressive Disclosure | No bundle reference files exist; the skill is a single 284-line file with the IaC destroy-command table and loop-interval guidance inlined rather than split into well-signaled one-level-deep references. | 2 / 3 |
Total | 10 / 12 Passed |