Content
70%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 an exceptionally rigorous, concrete operational playbook with best-in-class workflow sequencing and validation gates. Its weaknesses are token redundancy from duplicated graphviz edge lists and repeated warnings, and a broken reference layer: the three dispatch template files it depends on are not present in the bundle.
Suggestions
Add the missing implementer-prompt.md, task-reviewer-prompt.md, and re-review-prompt.md to the bundle (or inline their essential contracts), since the dispatch and review steps depend on them.
Cut the dot-graph duplication by declaring each node once (edge statements alone, or node declarations alone), and consolidate warnings stated multiple times (BASE vs HEAD~1, pre-judging, ledger survival) into a single statement each.
Consider moving the Model Selection detail and the Common Rationalizations table into a reference file to slim the always-loaded SKILL.md body.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is dense with novel, non-obvious operational guidance and never explains concepts Claude already knows, but it could be tightened significantly: both dot-graphs duplicate every node in the declaration and again in the edge list (~60 redundant lines), and several warnings repeat near-verbatim ("never HEAD~1, which silently drops..." three times, the pre-judging warning twice, the re-dispatched-tasks anecdote twice). This is more than the 'minor instances of over-explanation' of anchor 4. | 3 / 5 |
Actionability | Guidance is largely copy-paste ready: exact script invocations ("bash scripts/sdd-workspace PLAN_FILE", "bash scripts/review-package PLAN_FILE BASE HEAD"), exact ledger line formats, a four-status report taxonomy, and a worked example workflow. It falls short of anchor 5 because the three dispatch templates are load-bearing ('it is in the implementer template') yet absent from the bundle, leaving the dispatch steps not fully executable as written. | 4 / 5 |
Workflow Clarity | The multi-step process is explicitly sequenced with validation checkpoints and feedback loops throughout: a dual spec-plus-quality gate per task, a five-round fix loop with a scoped re-review each round, a round-5 breaker with adjudication rules, pre-flight conflict scan, and a final whole-branch review — plus a concrete example workflow. This matches the top anchor. | 5 / 5 |
Progressive Disclosure | Sections are well-organized and references are clearly signaled (templates linked inline, scripts named with paths, and the three scripts do exist in ./scripts/), but the three referenced template files — implementer-prompt.md, task-reviewer-prompt.md, re-review-prompt.md — are missing from the bundle, so navigation fails for the skill's core dispatch artifacts. Scored against the actual bundle structure per the rubric guideline, this is more than the 'minor organization gaps' of anchor 4. | 3 / 5 |
Total | 15 / 20 Passed |