Awesome-omni-skill scaffold-bulk-review-prototypes
Review all prototypes at once for cross-prototype consistency, coverage gaps, ADR follow-through, and scope discipline. Use for a full audit of all prototypes.
install
source · Clone the upstream repo
git clone https://github.com/diegosouzapw/awesome-omni-skill
Claude Code · Install into ~/.claude/skills/
T=$(mktemp -d) && git clone --depth=1 https://github.com/diegosouzapw/awesome-omni-skill "$T" && mkdir -p ~/.claude/skills && cp -r "$T/skills/design/scaffold-bulk-review-prototypes" ~/.claude/skills/diegosouzapw-awesome-omni-skill-scaffold-bulk-review-prototypes && rm -rf "$T"
manifest:
skills/design/scaffold-bulk-review-prototypes/SKILL.mdsource content
Bulk Prototype Review
Review every registered prototype for completeness, quality, and cross-prototype consistency.
Steps
1. Gather All Prototypes
- Read the prototypes index at
.scaffold/prototypes/_index.md - Read every prototype file in
(Globscaffold/prototypes/
).scaffold/prototypes/PROTO-*.md - Read the design doc at
.scaffold/design/design-doc.md - Read all system designs — Glob
.scaffold/design/systems/SYS-*.md - Read all engine docs from
.scaffold/engine/ - Read all ADRs — Glob
.scaffold/decisions/ADR-*.md - Read known issues at
.scaffold/decisions/known-issues.md - Read design debt at
.scaffold/decisions/design-debt.md - If no prototypes exist, report that and stop.
2. Per-Prototype Completeness
For each prototype, check pre-spike sections (all prototypes) and post-spike sections (Complete prototypes only):
Pre-Spike:
| Section | What "Complete" Means |
|---|---|
| Question | ONE specific, answerable question |
| Hypothesis | Expected answer with reasoning |
| Scope: Build | Numbered list of minimum steps |
| Scope: Skip | At least 2 explicit exclusions |
| Approach | Numbered execution steps |
| Related Documents | At least 1 valid scaffold doc reference |
Post-Spike (if Status is Complete):
| Section | What "Complete" Means |
|---|---|
| Answer | Direct answer to the Question |
| Evidence | Specific measurements or observations |
| Surprises | Findings listed or explicitly "None" |
| Design Impact | Table of affected docs with changes |
| ADRs Filed | Every Design Impact has a corresponding ADR |
| Disposition | Discarded / Archived / Absorbed with reason |
3. Per-Prototype Quality
For each prototype, check:
- Question is specific and singular — not compound or vague.
- Scope is minimal — Build scope is the minimum to answer the question.
- Skip boundaries are explicit — clear discipline boundary.
- Right-sized — achievable in one spike session.
4. Cross-Prototype Consistency
This is unique to bulk review — check relationships BETWEEN prototypes:
- Question overlap. Flag prototypes with substantially similar questions. Suggest merging or clarifying the distinction.
- Scope overlap. Flag prototypes with overlapping Build scopes — they may be doing redundant work.
- Coverage gaps. Cross-reference with systems, engine docs, slices, and known issues. Flag areas of significant uncertainty that have no prototype.
- ADR follow-through. For Complete prototypes: verify every Design Impact entry has a filed ADR. Flag missing ADRs.
- Disposition consistency. Flag Absorbed prototypes without a referenced TASK-###. Flag prototypes marked Complete but with empty post-spike sections.
- Stale prototypes. Flag Draft prototypes that reference documents which have since changed significantly or been deprecated.
- Status consistency. Check that index status matches file status for every prototype.
5. Registration Check
- Every prototype in
must have a corresponding file.scaffold/prototypes/_index.md - Every prototype file must be registered in
.scaffold/prototypes/_index.md - IDs, names, statuses, and dispositions must match between files and index.
Output Format
## Bulk Prototype Review — X prototypes audited ### Overview | ID | Name | Status | Pre-Spike | Post-Spike | Quality Issues | ADRs | |----|------|--------|-----------|------------|----------------|------| | PROTO-001 | ... | Draft | 5/6 | — | 1 issue | — | | PROTO-002 | ... | Complete | 6/6 | 5/6 | 0 issues | 2 | ### Per-Prototype Details #### PROTO-001 — [Name] | Section | Status | Notes | |---------|--------|-------| | ... | ... | ... | Quality Issues: - [specific issues with quotes] (repeat for each prototype) ### Cross-Prototype Consistency - **Question overlap:** [OK / overlapping prototypes found] - **Scope overlap:** [OK / redundant Build scopes found] - **Coverage gaps:** [OK / areas without prototypes] - **ADR follow-through:** [OK / X missing ADRs] - **Disposition consistency:** [OK / issues found] - **Stale prototypes:** [OK / X stale prototypes] ### Registration [Any mismatches between index and files] ### Recommendations (prioritized) 1. [Most impactful fix across all prototypes] 2. [Second priority] 3. ...
Rules
- This skill is read-only. Do not modify any files.
- Cross-prototype checks are the main value of bulk review over individual
calls. Emphasize relationship issues./scaffold-review-prototype - Be specific. Quote problematic text when flagging issues.
- Coverage gaps are high-value findings. Identifying what SHOULD be prototyped but isn't is as important as auditing existing prototypes.
- If all prototypes are well-written, say so. Don't manufacture issues.
- Prioritize recommendations by blast radius — issues that affect multiple prototypes or the pipeline rank higher.