Content
68%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 solid, command-heavy guide with strong executable CLI coverage for connection setup, action discovery/execution, and proxy fallback, plus explicit connection-state handling. Its main defects are a dangling 'Step 2' reference that breaks the workflow's step navigation, a never-instantiated action-run example, and light intro padding.
Suggestions
Define the steps explicitly (e.g. '## Step 1: Connect', '## Step 2: Search actions', '## Step 3: Run actions') so the two references to 'Step 2' resolve to a real section.
Add one concrete worked example for running an action, e.g. membrane action run list-candidates --connectionId=<id> --input '{"limit": 10}' --json, to replace the generic '{"key": "value"}' placeholder.
Trim the opening marketing paragraph ('Ashby is a recruiting software platform used by high-growth companies…') and the filler line 'Use action names and parameters as needed.' — they add tokens without operational value.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is mostly lean executable command blocks and tables, but contains trimmable padding: the marketing-style intro ('Ashby is a recruiting software platform used by high-growth companies…') and the filler line 'Use action names and parameters as needed.' Matches anchor 4 — efficient with minor instances of over-explanation. Not anchor 5: those extraneous sentences and some redundant auth explanations cost tokens without adding operational value. | 4 / 5 |
Actionability | Install, login, 'membrane connection ensure "https://www.ashbyhq.com/" --json', polling, action list/run, and the proxy flag table are concrete and executable, but '<actionId>' and '--input '{"key": "value"}'' are never instantiated with a real worked example (e.g. running 'list-candidates' from the table). Matches anchor 4 — mostly executable with minor gaps; not anchor 5 because no concrete end-to-end example covers a common case. | 4 / 5 |
Workflow Clarity | A sequence is implied by section order (Install → Authenticate → Connect → Search → Run) and there are real state checkpoints (READY / CLIENT_ACTION_REQUIRED / CONFIGURATION_ERROR with a re-poll loop), but the text says 'skip to **Step 2**' twice while no 'Step 2' is ever defined — a dangling reference in the core workflow. Matches anchor 3 (steps present but checkpoints/coherence gaps); not anchor 4 because broken step navigation is more than a minor validation gap. | 3 / 5 |
Progressive Disclosure | A single ~170-line file with no bundle files; sections are well-organized with clear headers, and the inlined 20-row actions table is reasonable for a single-connection CLI skill. No buried or nested references exist. Matches anchor 4 — good structure, content appropriately placed, minor gaps (the full actions table could live in a reference file); not anchor 5 because the overview/actions split could be cleaner. | 4 / 5 |
Total | 15 / 20 Passed |