Audit Api Usage logo

Audit Api Usage

Community
TomGranot
audit-api-usage

Inventory the integrations, service keys, legacy private apps, and internal tooling that call HubSpot APIs, and flag anything on legacy v1-v4 endpoints or legacy app architecture ahead of HubSpot's 2027 cutoffs. Produces a migration checklist to date-based API versions and service keys.

Overview

PublisherTomGranot
Repositoryhubspot-admin-skills
Skill nameaudit-api-usage
Stars
79
Forks
18
Bundled files
Instructions only
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by TomGranot on GitHub. Read the source before you install it.

Installation

Install the Audit Api Usage AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/TomGranot/hubspot-admin-skills/tree/main/skills/audit-api-usage
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/TomGranot/hubspot-admin-skills.git /tmp/hubspot-admin-skills
mkdir -p .claude/skills
cp -r /tmp/hubspot-admin-skills/skills/audit-api-usage .claude/skills/audit-api-usage
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Audit Api Usage in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Audit Api Usage on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Audit Api Usage is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Audit API Usage Before the 2027 Cutoffs

In March 2026 HubSpot replaced semantic API versioning (v1-v4) with date-based versions (/2026-09/-style paths, two releases a year, at least 18 months of support each). Two clocks are now running: v4 loses support on March 30, 2027, and v1-v3 are already unsupported, with enforcement in September 2027. Separately, HubSpot is retiring legacy private apps in favor of service keys. This skill inventories what in your stack calls HubSpot and produces a migration checklist for both.

Why This Matters

The deadline fails quietly: unsupported APIs "no longer receive updates, bug fixes, or stability guarantees," and marketplace apps that don't migrate risk losing certification. A portal typically has more callers than anyone remembers — service keys, legacy private apps, marketplace apps, middleware (Zapier/Make), form embeds, internal scripts, data warehouse syncs. Finding them takes an afternoon now or an incident later.

Key Facts

FactDetail
Current version2026-09 (for example /crm/objects/2026-09/contacts)
Release cadenceMarch and September, every year; each version is immutable
Support lifecycleAt least 18 months per version
v4 end of supportMarch 30, 2027
v1-v3 statusUnsupported since September 15, 2026 (no bug fixes, reliability, or security updates); enforcement September 2027
Legacy private appsCreation disabled October 26, 2026 for existing accounts (September 28, 2026 for new ones). Existing apps keep working and must migrate by September 2027
Replacement for private appsService keys (Settings > Integrations > Service Keys): bearer tokens for REST calls, no webhooks
Also deprecatedLegacy v1 OAuth API (token issuance/introspection)
Behavior change in 2026-09CRM writes enforce admin-configured required and conditionally required properties, returning HTTP 400 where v3 accepted the write

Sources: HubSpot developer changelog, "Introducing date-based API versioning" (2026-03-30), "Deprecating support for HubSpot v4 APIs" (2026-05-21), "Legacy private app creation sunset" (2026-08-27), and "Legacy APIs and legacy apps: what's going unsupported and when" (2026-09-15). Re-check these before quoting dates to stakeholders; HubSpot has moved sunset dates before.

Prerequisites

  • Super Admin access (to view Integrations settings and app logs)
  • Optionally: a service key or existing legacy private app token (HUBSPOT_ACCESS_TOKEN in .env) for the account-level usage query
  • Access to the source code of any internal integrations

Execution Pattern

Stage 1: Plan

Confirm with the user: which systems are known to touch HubSpot (CRM syncs, website forms, analytics pipelines, internal scripts), and who owns each.

Stage 2: Before — Build the Caller Inventory

Work through each discovery source and record every caller in a checklist (owner, what it does, endpoints/versions used):

  1. Legacy private apps and service keys: Settings > Integrations > Private Apps, and Settings > Integrations > Service Keys. Open each legacy app > Logs: the API call log shows the exact request paths (/crm/v3/..., /contacts/v1/...), which reveal the version in use. Record every legacy private app as a migration item in its own right, whatever endpoints it calls.
  2. Connected/marketplace apps: Settings > Integrations > Connected Apps. You can't see their internals; note each vendor — certified apps are being pushed to migrate by HubSpot, but confirm with the vendor for anything business-critical.
  3. Internal codebases: grep for version-revealing patterns:
    bash
    grep -rEn "api\.hubapi\.com/[a-z-]+/v[0-9]|/contacts/v1|/email/public/v1|hubapi.com/automation/v[23]" .
  4. Middleware (Zapier, Make, Workato, n8n): platform-managed connectors migrate on the vendor's schedule; custom HTTP steps inside them do not — check those by hand.
  5. Account-wide API usage (optional script check): the account-info API reports daily API usage totals for privately distributed apps:
    python
    resp = requests.get(f"{BASE}/account-info/v3/api-usage/daily", headers=HEADERS)
    This confirms how much traffic flows, not which versions — use the per-app logs from step 1 for version detail.

Stage 3: Execute — Classify and Plan Migrations

For every caller, classify:

StatusMeaningAction
RedCalls v1/v2 legacy endpoints (e.g., /contacts/v1/, /email/public/v1/, forms v2) or legacy OAuth v1Migrate now. Many of these have already been shut off (Contact Lists v1 returns 404 since April 30, 2026; Pipelines v1 sunsets December 4, 2026)
Amber (v4)Calls /v4/ endpoints, such as /automation/v4/flows or /crm/v4/ associationsSupported until 2027-03-30; migrate to date-based paths before then
Amber (v3)Calls /v3/ endpointsAlready unsupported, still working; migrate to 2026-09 date-based paths before September 2027
Amber (app)Runs on a legacy private app token, whatever it callsMove to a service key before September 2027
GreenCalls date-based paths with a service key or a projects-based app, or vendor-managed and confirmed migratingMonitor

For amber/red internal code, the migration is usually mechanical: same resources, new path shape (/crm/v3/objects/contacts becomes /crm/objects/2026-09/contacts) and sometimes renamed fields. HubSpot's migration guide says many v3 contracts carry over unchanged but not all, so test each endpoint rather than assuming. Batch the work per codebase, not per endpoint.

Test writes in particular. From 2026-09, a create or update that omits a property an admin marked required is rejected with MISSING_REQUIRED_PROPERTY or MISSING_CONDITIONAL_REQUIRED_PROPERTY. An integration that wrote sparse records through v3 will start failing after migration unless it sends those properties.

This repo's own scripts are amber: they target /crm/v3/, /marketing/v3/, and /automation/v4/. They work today, and they are on the migration list in CONTRIBUTING.md like everyone else's.

Stage 4: After

  1. Every caller in the inventory has a status and an owner.
  2. Red items have migration tickets with dates; v4 callers are scheduled before 2027-03, v3 callers and legacy private apps before 2027-09.
  3. Re-run this audit every 6 months — each March/September release starts a new support clock.

Rollback

Read-only — nothing to roll back.

Technical Gotchas

  1. App logs are the ground truth. Code greps miss dynamically-built URLs; the per-app request logs in Settings show what is actually being called.
  2. 404 vs 403 vs unsupported. After end of support, endpoints may keep working for a while — "unsupported" means no fixes or guarantees, not an instant shutoff. Do not treat "it still works" as "we're fine."
  3. You cannot create a new private app as a stopgap. After October 26, 2026 the option is gone for existing accounts. A new internal integration needs a service key from day one.
  4. Service keys do not receive webhooks. An integration that relies on webhook subscriptions needs a projects-based app, not a service key.
  5. OAuth is part of this. Apps using legacy v1 OAuth token endpoints must move to the date-based OAuth API even if their data endpoints are current.
  6. SDK versions pin API versions. If an integration uses an official HubSpot SDK, the SDK's major version determines which API version it calls — check the SDK changelog, not just your code.

Frequently asked questions

What does the Audit Api Usage AI skill do?

Inventory the integrations, service keys, legacy private apps, and internal tooling that call HubSpot APIs, and flag anything on legacy v1-v4 endpoints or legacy app architecture ahead of HubSpot's 2027 cutoffs. Produces a migration checklist to date-based API versions and service keys.

Why use Audit Api Usage on TypingMind?

Because you install it once and use it with any model. Audit Api Usage is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Audit Api Usage in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/TomGranot/hubspot-admin-skills/tree/main/skills/audit-api-usage. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Audit Api Usage?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Audit Api Usage?

As many as you like. As long as a model supports skills, you can use Audit Api Usage with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Audit Api Usage AI skill free?

Yes. It is published on GitHub by TomGranot under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇