.claude/skills/hld-review/SKILL.md
Evaluate an HLD note against linked ADRs, requirements, and architecture principles across six dimensions
npx skillsauth add DavidROliverBA/ArchitectKB hld-reviewInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
Evaluate a High-Level Design (HLD) note against linked ADRs, requirements, and architecture principles. Produces a structured review with scoring across six dimensions, severity-rated findings, and actionable next steps.
/hld-review <HLD note title>
/hld-review HLD - Remote Vault Access Architecture
/hld-review HLD - SecureTransfer Azure Files to S3 Mirroring
HLD - *.md file in the vault root.| Dimension | What Is Checked |
|---|---|
| ADR Coverage | Does each significant design choice have a corresponding ADR? Are referenced ADRs accepted? |
| Requirements Traceability | Are all referenced requirements (from relatedTo, project notes, linked concepts) addressed in the design? |
| Principle Compliance | Does the design align with vault architecture principles (Concept notes with conceptType: principle)? |
| Security Architecture | Are threat mitigations documented? Are security boundaries, encryption, and access controls explicit? |
| Integration Clarity | Are system interfaces, data flows, protocols, and error handling paths clearly defined? |
| Operational Readiness | Are monitoring, deployment strategy, runbook considerations, and rollback plans addressed? |
$ARGUMENTSHLD - *.md filesrelatedTo wiki-links from frontmattertags for domain contextconfidence, freshness, verified quality indicators[[...]] references)Read linked notes — for each wiki-link found in relatedTo and body text:
Query for architecture principles — find all Concept notes with conceptType: principle:
node .claude/scripts/graph-query.js --type=Concept --tag=conceptType/principle
If the graph query returns no results, fall back to:
grep -rl "conceptType: principle" *.md
Read each principle note to extract:
statement — the principle itselfimplications — what follows from the principledomain — which domain it applies toQuery for related threats — search for Threat notes relevant to the HLD's domain:
domain/security, domain/cloud, domain/integration)node .claude/scripts/graph-query.js --type=ThreataffectedSystems and domain overlapCheck for related ADRs not yet linked — search for ADRs whose relatedTo or project fields reference the same systems or projects as the HLD:
node .claude/scripts/graph-query.js --type=Adr --tag=status/accepted
Score each dimension as PASS, PARTIAL, or FAIL using the criteria below.
| Score | Criteria | |---|---| | PASS | Every significant design choice (technology selection, architectural pattern, integration approach) has a corresponding accepted ADR | | PARTIAL | Some design choices have ADRs but others are undocumented, or ADRs exist but are still in draft/proposed status | | FAIL | Major design choices lack ADR coverage entirely |
How to assess:
relatedTo links or vault search)draft or proposed rather than accepted| Score | Criteria | |---|---| | PASS | All requirements referenced in linked project/concept notes are explicitly addressed in the HLD | | PARTIAL | Most requirements are addressed but some are missing or only implicitly covered | | FAIL | Significant requirements gaps — linked requirements not addressed in the design |
How to assess:
| Score | Criteria | |---|---| | PASS | Design explicitly aligns with all applicable architecture principles; no contradictions | | PARTIAL | Design is broadly consistent but some principles are not explicitly addressed | | FAIL | Design contradicts one or more architecture principles, or ignores critical principles |
How to assess:
statement and implicationsdomain matches the HLD's domain tags| Score | Criteria | |---|---| | PASS | Security boundaries, encryption (in-transit and at-rest), authentication, authorisation, and threat mitigations are all explicitly documented | | PARTIAL | Some security aspects are covered but gaps exist (e.g., encryption mentioned but no auth model) | | FAIL | Security is not meaningfully addressed in the HLD |
How to assess:
| Score | Criteria | |---|---| | PASS | All system interfaces are documented with protocols, data formats, error handling, and SLAs | | PARTIAL | Interfaces are identified but details are incomplete (missing protocols, no error handling) | | FAIL | System interfaces are vague or missing — unclear how components communicate |
How to assess:
| Score | Criteria | |---|---| | PASS | Monitoring, alerting, deployment strategy, rollback plan, and operational runbook considerations are documented | | PARTIAL | Some operational aspects are covered but the design lacks deployment or monitoring detail | | FAIL | No operational considerations — the HLD focuses only on functional architecture |
How to assess:
Categorise each finding by severity:
| Severity | Definition | Action Required | |---|---|---| | BLOCKING | Fundamental gap that must be resolved before the design can proceed to implementation | Rework required | | ADVISORY | Gap that should be addressed but does not prevent progress | Address before go-live | | INFORMATIONAL | Observation or suggestion for improvement | Consider in future iterations |
Based on the dimension scores and finding severities:
| Status | When Applied | |---|---| | APPROVED | All dimensions PASS, no BLOCKING findings | | APPROVED WITH CONDITIONS | No dimensions FAIL, but PARTIAL scores exist or ADVISORY findings need attention | | NEEDS REWORK | Any dimension scores FAIL, or any BLOCKING findings exist |
Output the review directly to the console. Use the template below.
# HLD Review: [HLD Title]
**Reviewed:** [today's date]
**HLD Created:** [created date from frontmatter]
**Last Reviewed:** [reviewed date from frontmatter]
**Confidence:** [confidence from frontmatter]
---
## Review Status: [APPROVED / APPROVED WITH CONDITIONS / NEEDS REWORK]
---
## Scoring Summary
| Dimension | Score | Evidence | Gaps |
|---|---|---|---|
| ADR Coverage | [PASS/PARTIAL/FAIL] | [what ADRs exist and cover] | [what design choices lack ADRs] |
| Requirements Traceability | [PASS/PARTIAL/FAIL] | [requirements addressed] | [requirements not addressed] |
| Principle Compliance | [PASS/PARTIAL/FAIL] | [principles aligned with] | [principles not addressed or contradicted] |
| Security Architecture | [PASS/PARTIAL/FAIL] | [security measures documented] | [security aspects missing] |
| Integration Clarity | [PASS/PARTIAL/FAIL] | [interfaces documented] | [interfaces lacking detail] |
| Operational Readiness | [PASS/PARTIAL/FAIL] | [operational aspects covered] | [operational aspects missing] |
---
## Findings
### BLOCKING
[If none: "No blocking findings."]
1. **[Finding Title]** — [Description of the gap, which dimension it affects, and why it blocks progress]
- **Dimension:** [affected dimension]
- **Recommendation:** [specific action to resolve]
### ADVISORY
[If none: "No advisory findings."]
1. **[Finding Title]** — [Description and recommendation]
- **Dimension:** [affected dimension]
- **Recommendation:** [specific action to resolve]
### INFORMATIONAL
[If none: "No informational findings."]
1. **[Finding Title]** — [Observation or suggestion]
---
## Conditions
[Only include if status is APPROVED WITH CONDITIONS]
The following conditions must be addressed before implementation:
- [ ] [Condition 1 — specific, actionable item]
- [ ] [Condition 2]
---
## Context Analysed
### Linked ADRs
| ADR | Status | Relevance |
|-----|--------|-----------|
| [[ADR - Title]] | accepted | [how it relates to the HLD] |
### Architecture Principles Assessed
| Principle | Domain | Compliance |
|-----------|--------|------------|
| [[Concept - Principle Name]] | [domain] | [aligned/not addressed/contradicted] |
### Threat Coverage
| Threat | Severity | Mitigated? |
|--------|----------|------------|
| [[Threat - Name]] | [severity] | [yes/partial/no] |
### Systems Referenced
[List of System notes referenced in the HLD with their status]
---
## Next Steps
- [ ] [Action item based on findings — e.g., "Create ADR for Docker orchestration choice"]
- [ ] [Action item — e.g., "Add monitoring section to HLD"]
- [ ] [Action item — e.g., "Link Threat - Denial of Service and document mitigation"]
- [ ] Update HLD `reviewed` date to today
- [ ] Update HLD `verified` to `true` once conditions are met
/hld-review HLD - Remote Vault Access Architecture
Reviews the HLD against all six dimensions, checking linked ADRs, vault principles, relevant threats, and operational completeness.
/hld-review HLD - SecureTransfer Azure Files to S3 Mirroring
Useful after new ADRs are accepted to verify the HLD still aligns with the latest decisions.
relatedTo field — this is itself a finding (ADVISORY: "HLD lacks cross-references to ADRs, systems, and requirements")/nfr-review — Review an HLD against NFR requirements (complementary to this skill)/diagram-review — Analyse architecture diagrams for readability and quality/adr — Create new ADRs for design choices identified as gaps/diagram — Generate architecture diagrams referenced in the HLD/impact-analysis — Analyse the impact of changes to systems in the HLD.claude/rules/naming-conventions.md — HLD naming patterns.claude/context/frontmatter-reference.md — HLD frontmatter schema.claude/context/architecture.md — Architecture governance contexttools
--- context: fork --- # /youtube Save a YouTube video as both a Weblink (quick reference) and a detailed Page (full analysis). ## Usage ``` /youtube <url> /youtube <url> <optional title override> ``` ## Examples ``` /youtube https://www.youtube.com/watch?v=0TpON5T-Sw4 /youtube https://youtu.be/abc123 AWS re:Invent Keynote ``` ## Prerequisites This skill uses the MCP Docker YouTube tools: - `mcp__MCP_DOCKER__get_video_info` - Video metadata - `mcp__MCP_DOCKER__get_transcript` - Full trans
data-ai
Create and manage git worktrees for parallel agent sessions
testing
--- context: fork --- # /wipe Generate a context handoff summary, clear the session, and resume in a fresh conversation. Detects environment and provides automated (tmux) or manual workflow. ## Usage ``` /wipe /wipe quick # Minimal handoff, just essentials /wipe detailed # Comprehensive handoff with full context ``` ## Instructions When the user invokes `/wipe`: ### Phase 1: Detect Environment First, check the terminal environment: ```bash echo "Environment Detection:"
data-ai
--- context: fork --- # /weekly-summary Generate comprehensive weekly summary from daily notes, meetings, tasks, and project updates using parallel sub-agents. ## Usage ``` /weekly-summary /weekly-summary --last-week /weekly-summary --from 2026-01-01 --to 2026-01-07 /weekly-summary --output page # Create Page note instead of just outputting ``` ## Instructions This skill uses **5 parallel sub-agents** to gather data concurrently from different vault areas, then synthesizes a comprehensi