Skip to content
Mobile
Agent

data-integrity

Read before assistant tools, WebLLM configuration, or ZoneMinder schemas.

BOOST
From plugin
zmninjang
1089 skills9 agents1 command
Install
$ npx -y skills add ZoneMinder/zmNinjaNg --agent claude-code

How it fires

How this agent gets triggered: by you, by Claude, or both.

  • Fires itselfAuto-invocation. Claude auto-loads it when your prompt matches the work.Auto-invocation is when the right skill fires by itself at the right moment, driven by a FLOW.md router and a hook, instead of you invoking it by name. It is the difference between a skill being installed and a skill actually getting used.Read the full definition →
  • You can call itInvoke it directly when you want it.

Context preview

The summary Claude sees to decide when to auto-load this agent.

Read before assistant tools, WebLLM configuration, or ZoneMinder schemas.

Agent definition

data-integrity.md

Data integrity playbook

Read before assistant tools, WebLLM configuration, or ZoneMinder schemas.

  • A library's merged configuration is unknown until runtime verifies it. Absence in one source layer proves nothing.
  • Treat every model tool argument as untrusted. Validate at `execute` entry, name valid values on failure, then query.
  • ZoneMinder fields drift. Schemas must use `withFieldCatch(shape, identity)`; arrays use `tolerantArray(itemSchema, label)`.
  • Keep unknown response fields stripped, never strict. ZoneMinder-controlled vocabularies use `z.string()`, not `z.enum()`.
  • A row failing identity drops. A field failing type falls back to a type-matching default. Remain strict for values sent to ZoneMinder.
  • Update `api/__tests__/types.test.ts` when adding schemas.
Ships withzmninjang

Documentation A ZoneMinder client for web, desktop, iOS, and Android. It shows live camera feeds, a montage of many cameras at once, and events on a timeline, and has an AI chat agent you can ask about your cameras.

Get the whole plugin

Other agents on zmninjang.

glossary
Agent

glossary

One name per concept, for code identifiers, briefs, docs, and commit messages. Each entry says what the thing is, not what it…