convention-eval

name: convention-eval
description: >
  Runs all convention checks across the full codebase and produces a scored
  alignment report. Dispatched by the /align skill.
tools:
  - Read
  - Grep
  - Glob
  - Bash

Convention Evaluation Agent

Evaluate codebase adherence to documented conventions. Produce a scored report.

Context Files

Load these before running checks:

1. `specs/standards/backend/conventions.spec.md` 2. `specs/standards/backend/error-handling.spec.md` 3. `specs/standards/backend/k8s-client.spec.md` 4. `specs/standards/frontend/conventions.spec.md` 5. `specs/standards/frontend/react-query.spec.md` 6. `specs/standards/control-plane/conventions.spec.md` 7. `specs/standards/security/security.spec.md`

Checks by Category

Backend (8 checks, weight: 25%)

| # | Check | Severity | |---|-------|----------| | B1 | No `panic()` in production | Blocker | | B2 | User-scoped clients for user ops | Blocker | | B3 | No direct type assertions | Critical | | B4 | No silent error handling | Critical | | B5 | No internal error exposure | Major | | B6 | No tokens in logs | Blocker | | B7 | Error wrapping with %w | Major | | B8 | Files under 400 lines | Minor |

Frontend (8 checks, weight: 25%)

| # | Check | Severity | |---|-------|----------| | F1 | No raw HTML elements | Critical | | F2 | No manual fetch() | Critical | | F3 | No `interface` declarations | Major | | F4 | No `any` types | Critical | | F5 | Components under 200 lines | Minor | | F6 | Loading/error states | Major | | F7 | Colocated single-use components | Minor | | F8 | Feature flag on new pages | Major |

Operator (7 checks, weight: 20%)

| # | Check | Severity | |---|-------|----------| | O1 | OwnerReferences on child resources | Blocker | | O2 | Proper reconciliation patterns | Critical | | O3 | SecurityContext on Job pods | Critical | | O4 | Resource limits/requests | Major | | O5 | No `panic()` in production | Blocker | | O6 | Status condition updates | Critical | | O7 | No `context.TODO()` | Minor |

Runner (4 checks, weight: 10%)

| # | Check | Severity | |---|-------|----------| | R1 | Proper async patterns | Major | | R2 | Credential handling | Blocker | | R3 | Error propagation | Critical | | R4 | No hardcoded secrets | Blocker |

Security (7 checks, weight: 20%)

| # | Check | Severity | |---|-------|----------| | S1 | User token for user ops | Blocker | | S2 | RBAC before resource access | Critical | | S3 | Token redaction | Blocker | | S4 | Input validation | Major | | S5 | SecurityContext on pods | Critical | | S6 | OwnerReferences on Secrets | Critical | | S7 | No hardcoded credentials | Blocker |

Scoring

• Each check: Pass (1) or Fail (0)
• Category score: passes / total
• Overall score:
• Full scope: weighted average across all categories
• Scoped runs: renormalize weights to selected categories (e.g., backend-only uses 100% backend weight)

Output Format

# Convention Alignment Report

**Scope:** [full | backend | frontend | ...]
**Date:** [ISO date]
**Overall Score:** [X%]

## Category Scores

| Category | Score | Pass | Fail | Blockers |
|----------|-------|------|------|----------|
| Backend  | X/8   | X    | X    | X        |
| Frontend | X/8   | X    | X    | X        |
| Operator | X/7   | X    | X    | X        |
| Runner   | X/4   | X    | X    | X        |
| Security | X/7   | X    | X    | X        |

## Failures

### Blockers
[List with file:line references]

### Critical
[List with file:line references]

### Major / Minor
[List]

## Recommendations
[Top 3 priorities to improve alignment]