Content
96%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 strong, highly actionable body: concrete executable commands with flags and worked examples, a clearly sequenced setup-to-diagnostics workflow with status validation and error-recovery loops, and a well-signaled reference file. The only weaknesses are localized redundancy (the setup-issues table duplicates the Error Reference) and some reference-level detail (output columns, format flags) that could live in references/commands.md.
Suggestions
Merge the 'Common setup issues' table with the 'Error Reference' table — they list the same six symptom/fix pairs almost verbatim — and keep a single error section.
Move the per-command output-columns list and the output-format (-f) table into references/commands.md, leaving only a pointer in the body to slim the main file.
Fold Step 5 Diagnostics into the error reference (or the status rule in Step 3), since it repeats the status/launch guidance already given in Steps 1 and 3.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is dense command tables and rules with no explanations of concepts Claude already knows, but the 'Common setup issues' table (6 rows) is duplicated nearly verbatim in the 'Error Reference' table, and Step 5 Diagnostics repeats the status/launch guidance from Steps 1 and 3. Matches 'efficient; minor instances of over-explanation that could be trimmed'; not 5 due to the duplicated error tables, not 3 since the padding is localized rather than pervasive. | 4 / 5 |
Actionability | Fully executable copy-paste commands with real flags and concrete instances ('opencli tradingview options-chain --ticker SNDK --expiry 2026-05-22 -f json', 'quote --ticker SPY --exchange NYSEARCA'), enumerated output columns per command, and an error-to-fix table. Matches the 5 anchor — specific examples cover the common cases. | 5 / 5 |
Workflow Clarity | Steps 1-5 are clearly sequenced (status check with READY/SETUP_NEEDED/NOT_INSTALLED branching, install, identify need, execute, present, diagnostics) with explicit validation checkpoints and feedback loops (error reference with fixes, 'Wait a few seconds, then retry', relaunch via `opencli tradingview launch`). Read-only skill, so the destructive-operation cap does not apply; matches the 5 anchor's explicit validation and error-recovery loops. | 5 / 5 |
Progressive Disclosure | The body is an overview with one clearly signaled, one-level-deep reference (`references/commands.md`, a real file, cited both inline — 'See references/commands.md for the operations cheat sheet' — and in a Reference Files section). Matches 'good structure; most content appropriately placed; minor organization gaps'; not 5 because reference-grade detail (full output-columns list and format-flag table) is inlined in the body, and only one bundle file exists to absorb it. | 4 / 5 |
Total | 18 / 20 Passed |