From spectre
Conducts principal architecture review of completed features, focusing on compounding decisions, architectural debt, missed abstractions, and performance cliffs. Produces a structured report with actionable findings.
How this skill is triggered — by the user, by Claude, or both
Slash command
/spectre:spectre-architecture_reviewThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Treat the current command arguments as this workflow's input. When invoked from a slash command, use the forwarded `$ARGUMENTS` value.
Treat the current command arguments as this workflow's input. When invoked from a slash command, use the forwarded $ARGUMENTS value.
feature_description: Feature being reviewed — (via ARGUMENTS). Ifpaths_or_files: Relevant paths to examine — (via ARGUMENTS)arch_notes_or_docs: Existing architecture context — (via ARGUMENTS, optional)Review Philosophy (cross-cutting guidelines):
Constraints:
Adapt
paths_or_filesarch_notes_or_docs if providedOutput Format — Architecture review with these sections:
2-3 sentences: What's the architectural story here? Is this feature setting us up well or creating future pain?
Issues that will cause significant pain if left unaddressed—architectural violations, major performance problems, or patterns that will spread.
For each issue:
Places where we're overengineering, where a simpler approach would work, or where we've introduced abstractions we don't yet need.
Only mention performance if there's a clear problem or a clear win—don't speculate about micro-optimizations.
Briefly note architectural choices worth preserving or patterns worth spreading to other parts of the codebase.
Step 1 - Explore Implementation:
Step 2 - Produce Review Report:
npx claudepluginhub codename-inc/spectre --plugin spectrePerforms systematic architecture reviews across 7 dimensions (structural, scalability, enterprise readiness, performance, security, ops, data) with scored reports and prioritized recommendations.
Analyzes any codebase's architecture with 6 specialist agents (perf/scale, reliability, security, ops/DX, data/deps + Codex cross-review). Agents debate risks, fragile spots, improvements for audits/refactors.
Orchestrates parallel reviews from all architecture specialist subagents, aggregates into consolidated report for full multi-perspective assessments.