Content
88%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 a highly actionable, well-gated end-to-end workflow: every step has executable commands, explicit user-confirmation gates, polling/error-recovery loops, and post-migration application steps. Its only weaknesses are mild duplication of the frontmatter's framework list and reference links that are listed rather than contextually signposted.
Suggestions
Trim the supported-framework tables in the body (they duplicate the frontmatter description) to a brief pointer, keeping only the migration-vs-stays tables that add new information.
Replace the flat 'Required files' list with contextual references — e.g., note when the agent should consult references/launchwithaws-2026-06-15.json (the service model) or a specific script, so navigation is signaled rather than merely enumerated.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is efficient — bare bash commands, JSON response shapes, and compact tables instead of prose explanations of concepts Claude already knows — but the supported-framework tables largely duplicate the frontmatter description and a few passages (e.g., the repeated migration-vs-stays summary sentence) could be trimmed. It sits between 'efficient with minor trimming possible' (4) and 'every token earns its place' (5), closer to 4. | 4 / 5 |
Actionability | Every step provides copy-paste-ready commands ("python3 scripts/launch_with_aws.py auth-start", "curl -L -o /tmp/migration-snapshot.zip \"<download_url>\"", the rsync/git sequence in Step C), exact JSON outputs, and verbatim confirmation text to present to the user. This fully matches the anchor for executable, copy-paste-ready guidance covering the common cases. | 5 / 5 |
Workflow Clarity | The flow is a clearly sequenced 1–8 process with explicit validation checkpoints: confirmation gates ("Do NOT call create-launch until the user explicitly confirms"), status polling with progression (analyzing → planning → planned → executing → completed/failed) and sleep intervals, failureReason-based error recovery, a dirty-tree check before applying changes, and conflict-handling instructions. This matches the top anchor for explicit validation steps and error-recovery feedback loops. | 5 / 5 |
Progressive Disclosure | Structure is good: the SKILL.md body is an overview/workflow, implementation lives in the six scripts, and all referenced files (launch_with_aws.py, launch_config.py, auth.py, auth_callback_server.py, launch_api_client.py, archive.py, references/launchwithaws-2026-06-15.json) exist and are one level deep. Not 5 because the 'Required files' line lists references without contextual navigation — it never signals when or why the agent should read the service model JSON or the individual scripts. | 4 / 5 |
Total | 18 / 20 Passed |