Menu
">"
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Based on SOC occupation classification
| id | model-currency-checker |
| name | model-currency-checker |
| type | Capability Uplift |
| version | 1.0.0 |
| created | 20/03/2026 |
| modified | 20/03/2026 |
| status | active |
| triggers | ["check model versions","are models up to date","model currency","which models are we using","audit ai models","check ai versions","model audit","update models"] |
| description | > |
Purpose: Prevent the system from running on stale, deprecated, or unapproved AI models. Model capability compounds — running on outdated models degrades output quality silently.
Use this skill when:
pnpm starter:audit is run| Task Category | Provider | Approved Model ID | Approved Since |
|---|---|---|---|
| Reasoning / orchestration | Anthropic | claude-sonnet-4-6 | 06/03/2026 |
| Complex reasoning | gemini-2.5-pro-preview | 06/03/2026 | |
| Fast image gen / editing | gemini-2.5-flash-image | 06/03/2026 | |
| Image editing with context | Nano Banana | nano-banana-pro | 06/03/2026 |
| High-fidelity branding | imagen-4 | 06/03/2026 | |
| 3D logo renders | imagen-4 | 06/03/2026 |
Review cycle: Model policy must be reviewed every 90 days. Next review due: 06/06/2026
Search these file types for model ID strings:
*.ts, *.js — TypeScript/JavaScript source*.py — Python source*.json — Config files (model IDs in env or config)*.yaml, *.yml — CI/CD configs, docker compose*.env, *.env.example — Environment filesSearch patterns to find:
gemini- prefiximagen- prefixclaude- prefixgpt- prefix (should not be present — unapproved)ollama references (check version)For each model ID found:
| Classification | Meaning |
|---|---|
CURRENT | Model ID matches approved policy exactly |
OUTDATED | Model ID was previously approved but has been superseded |
UNAPPROVED | Model ID not in approved policy |
REVIEW_NEEDED | Model ID format matches but version cannot be confirmed |
Run pnpm starter:audit to generate reports/model-currency-report.md automatically.
Or manually: grep the codebase and compare against the table above.
For each OUTDATED or UNAPPROVED model:
MODEL CURRENCY REPORT
═══════════════════════════════════════════════════
Date: [DD/MM/YYYY]
Codebase: [root path]
APPROVED POLICY
─────────────────
[table of approved models]
FINDINGS
─────────────────
CURRENT: [file:line] — [model-id] — [category]
OUTDATED: [file:line] — [model-id] — Replace with: [approved-id]
UNAPPROVED: [file:line] — [model-id] — Not in approved policy
SUMMARY
─────────────────
Current: [N]
Outdated: [N]
Unapproved: [N]
ACTION REQUIRED: Update [N] model reference(s) before shipping.
═══════════════════════════════════════════════════
Before marking model check complete:
reports/model-currency-report.md| Failure | Recovery |
|---|---|
| No model references found | Verify scan covered the right file types and directories |
| Model ID found in env file but not code | Check if env var is actually used |
| Model ID in comment only | Mark as informational, not a live reference |
| Model has been deprecated by provider | Update to approved replacement immediately |
Finding: apps/backend/src/models/anthropic.py:14 — claude-2 — OUTDATED
Action: Replace with claude-sonnet-4-6
Verification: Test prompt round-trip after update
Finding: "I think the models are up to date." Reason rejected: No scan performed, no evidence, opinion not evidence.
Apply this skill for Unite-Hub Supabase migrations, PostgREST/Data API visibility, founder-scoped Playwright journeys, or errors such as PGRST205, access=denied, stale Supabase linked refs, or migration history drift. Prevents repeating the SQL/cache/auth loop by enforcing the exact verification sequence for core journeys.
The compass for Unite-Hub's road to /shipit. Defines the single NorthStar (a real, comprehensive, working founder CRM in production, every section GREEN), the binding definition of GREEN, and the No-Invaders Manifest that keeps the build honest and surgical. Consult BEFORE deciding what to build/skip/finish — it resolves "200 ≠ real" temptations and scope-creep pressure. P1, auto-loaded.
Apply this skill WHEN scaffolding a new cron "pull" route that syncs external/derived data into Supabase on a schedule (Vercel cron). Encodes the Unite-Hub cron invariants: CRON_SECRET auth, FOUNDER_USER_ID actor, overlap safety, idempotent upsert, last-sync timestamp, and failure surfacing. Generic `cron-scheduler` covers scheduling; this covers the PULL handler body. P3.
Apply this skill WHEN verifying that a route, page, or integration serves REAL data and not silent mock/placeholder data. Detects the "false-green" failure mode: an endpoint returns 200 (or a page renders) while the underlying data is fabricated because a provider is unconnected. Trigger WHENEVER classifying a section's readiness, reviewing integration wrappers, or before marking anything GREEN. P2 — load on audit/verify tasks.
Manifest-first context isolation — each subagent receives only its scope, never the full codebase
Apply this skill for ANY decision with non-obvious tradeoffs: architectural choices, debugging without a clear root cause, performance strategies, security decisions, feature design with competing constraints, refactoring scope decisions. Forces multi-perspective analysis before committing to a solution. P1 auto-load — always active on complex reasoning tasks.