Content
82%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 content is a strong, example-driven CLI reference: every section pairs precise behavioral semantics with executable commands, and failure/validation behavior is consistently surfaced. Weaknesses are minor — a few dense prose passages that could be tightened, and some reference-grade detail that could be split into bundle files now that the document exceeds a single-screen overview. No bundle files exist in references/, scripts/, or assets/, so progressive disclosure was scored on the body's own structure.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is information-dense with almost no explanation of concepts Claude already knows — every paragraph covers tool-specific semantics like "The PDB span is executable source, not a complete member boundary" and "PDBs *carry* SourceLink data; they are not SourceLink themselves". Not 5 because a few convoluted passages (e.g. the "For one package with exactly `SourceLink: Files` selected..." paragraph and the provenance-boundary sentences in the intro) could be tightened or trimmed. | 4 / 5 |
Actionability | Five command blocks give fully executable, copy-paste-ready invocations covering the common cases — `dnx dotnet-inspect -y -- library System.Text.Json -S "SourceLink: Files"`, `member JsonSerializer --platform System.Text.Json Serialize:1 -S "PDB Source"`, and `--print --row 1` variants — with concrete flags (--urls, --paths, --json-array) explained inline. Not 4 because the examples span locate/fetch/print workflows and specific packages rather than leaving minor gaps. | 5 / 5 |
Workflow Clarity | Sections follow a clear sequence — "Find where the source lives" → "Fetch PDB source" → "Inspect or print authored member parts" → "URL forms" — and validation behavior is described (checksum verification of fetched bodies, "Unavailable parts fail rather than selecting a substitute", missing printable documents are reported). Not 5 because the checkpoints are stated as tool behavior rather than explicit validate-then-proceed steps with error-recovery guidance; not 3 because the ordering and failure semantics are consistently made explicit. | 4 / 5 |
Progressive Disclosure | The body is well-organized into focused one-level sections with no nested references and no monolithic wall of text; cross-references to the complementary `decompiler` and `signals` skills are named with their purpose. Not 5 because the document is ~135 lines with detail-heavy passages (row-selection semantics, part-name lists, URL-form options) that could live in reference files, and the sibling-skill mentions lack paths or links. | 4 / 5 |
Total | 17 / 20 Passed |