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 highly actionable, well-sequenced agent prompt with copy-paste gh commands, explicit preconditions, an output checklist, and complete output specification. Its main cost is token duplication: the recurrence rule and the no-scripts/no-pipes rules are each stated two to three times across Step 2, Step 3/4, and Tool Guidance, which could be consolidated without losing clarity.
Suggestions
Consolidate the closed-issue recurrence guidance: state the open-first/recurrence-signal rule once (Step 2c or Step 3, not both) and reference it from the trend_direction field spec, saving ~15 lines.
Merge the no-script/no-pipe rules into the single Tool Guidance section and drop their restatements in Step 2c ("Do not write Python, Node, or shell scripts...") and Step 4 ("do not pipe through python3..."), keeping only the one hard-rule line per location if emphasis is needed.
The <examples> block and Integration Points section could move to the caller's documentation or a reference file to shorten the always-loaded prompt body.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body assumes Claude's competence (no concept explanations, no padding) but repeats the same guidance multiple times: the closed-issue recurrence rule appears in Step 2c ("How to interpret closed issues"), again as Step 3 clustering rule 1, and again in the trend_direction field spec; the no-scripts/no-pipes rule is stated in Step 2c, Step 4 ("do not pipe through python3"), and then twice more in the Tool Guidance section. This matches the 'mostly efficient but could be tightened' anchor rather than 4's 'minor instances', while not reaching 2's 'several unnecessary explanations or padded sections'. | 3 / 5 |
Actionability | Fully executable throughout: exact copy-paste gh commands with complete --jq filters (e.g., `gh issue list --state open --limit 100 ... --jq '[.[] | {number, title, labels, createdAt, body: (.body[:500])}]'`), exact precondition commands, verbatim fallback messages for failure cases, and a fully specified output template with field-level requirements. This matches the anchor-5 'copy-paste ready commands; specific examples cover the common cases'. | 5 / 5 |
Workflow Clarity | Steps 1-6 are clearly sequenced with explicit validation checkpoints: ordered precondition checks with stop-on-failure behavior ("If any fails, return a clear message explaining what is missing and stop"), an accuracy requirement section, and an output checklist ("Output checklist — verify before returning") that functions as an explicit pre-return validation with per-item checks. This matches the anchor-5 'explicit validation steps; feedback loops for error recovery; checklists for complex processes'. | 5 / 5 |
Progressive Disclosure | Scored against the actual bundle structure: no references/, scripts/, or assets/ directories exist, no nested references, and all sections are clearly organized (Methodology, Output Format, Tool Guidance, Integration Points). This matches anchor 4's 'good structure; most content is appropriately placed'. It is not 5 because a ~220-line body makes no use of the overview-plus-detail split the anchor-5 example demonstrates (and the under-50-line simple-skill exception does not apply); it is above 3 because no content that clearly belongs in a separate file is buried inline. | 4 / 5 |
Total | 17 / 20 Passed |