Content
77%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
The body is highly actionable with a clearly sequenced, gated workflow and concrete executable code, but it is somewhat redundant and monolithic for its length with no progressive disclosure.
Suggestions
Remove the redundant "Important" section or trim it to only constraints not already stated in Context Budget and the steps.
Factor the pinned output-format blocks (Repo Match Results, Resolved Repository) into a short reference snippet to reduce inline bulk, or at least de-duplicate the repeated format rules.
Consolidate the repeated "name + description only / no cloning to inspect" guidance into a single canonical statement.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Mostly concrete and action-oriented, but the closing "Important" section restates constraints already given in Context Budget and Step 3 ("Name + description only", "Local repos first"), so it could be tightened. | 2 / 3 |
Actionability | Provides executable bash for metadata scanning, lib calls, scoped remote fetching, and pinned machine-checked output blocks with a concrete example — copy-paste ready. | 3 / 3 |
Workflow Clarity | Six steps are clearly sequenced with explicit gates ("Only reach this step if no local repo matched confidently"), machine-checked output validation, and user confirmation before cloning. | 3 / 3 |
Progressive Disclosure | Well-organized into named sections, but it is a single ~200-line file with no one-level-deep reference split; the determinism-lib usage and output-format specs are inline rather than factored out. | 2 / 3 |
Total | 10 / 12 Passed |