Content
81%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.
A dense, highly actionable instruction body with an excellent validated workflow: probe-before-live checks, status confirmations, and recovery loops for expired logins. The main gaps are the absent server implementation code and repeated sender-key prohibitions that could be consolidated.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean and imperative with no explanations of concepts Claude already knows ('Bind the server to 0.0.0.0:<port>, not 127.0.0.1. Tailscale peers cannot reach a localhost-only bind.'). Not 5 because the sender-key prohibitions are repeated roughly five times across sections ('Do not put the sender key in the browser, in chat, or in this skill', 'The user must not paste the sender key in chat', 'Do not print the value. Do not log the value', 'Do not print the sender key, tokens, or cookies') — a minor redundancy that could be consolidated. | 4 / 5 |
Actionability | Mostly executable: exact update_state fields, a concrete secret-request card template, precise POST spec (headers, 8-second timeout, one try), and copy-paste commands ('curl -fsSL https://tailscale.com/install.sh | sudo sh', 'sudo tailscale up --hostname=<short-name> --accept-dns=false --ssh=false'). Not 5 because the central deliverable — the local server that forwards POSTs to the webhook — has only a requirements list with no code, leaving a gap in executable guidance. | 4 / 5 |
Workflow Clarity | The sequence (create routine → copy URL/key → request secret → host page → tailnet → handle wake) is clear and includes explicit validation with feedback loops: 'Before you tell the user that the UI is live, probe once with a harmless payload', 'confirm with tailscale status and tailscale ip -4', 'Probe http://<100.x.x.x>:<port>/ and expect HTTP 200', and recovery for expiry ('If the login URL expires, run tailscale up again and send the new URL'). | 5 / 5 |
Progressive Disclosure | The body is well-organized into clearly titled sections (Create the webhook routine, Copy the URL and the sender key, Host the page, Put the page on the tailnet, Handle the webhook wake) with one clearly signaled, one-level-deep external reference ('read [the Claude Code adapter](../../docs/harness/claude-code.md) once per session'). No bundle files exist to offload content into, and at ~110 lines with a coherent single-task flow the structure is good, though a brief overview of the end-to-end flow at the top would improve navigation. | 4 / 5 |
Total | 17 / 20 Passed |