Atlas / Skills / jeremylongshore / Ideogram Reference Architecture

Ideogram Reference ArchitectureSAFE

skills/jeremylongshore/ideogram-reference-architecture

Model-agnostic agent-skills platform with a harness-free canonical layer, verified adapters, and the ccpi package manager. Explore at tonsofskills.com.

Verdict
SAFE
Grade
B
Trust score
89 /100
Version
1.11.0
Hosts
1 documented
License
MIT
Stars
2,824
01

Overview

Model-agnostic agent-skills platform with a harness-free canonical layer, verified adapters, and the ccpi package manager. Explore at tonsofskills.com.

Read from source at commit 4f83675ca38aOBSERVED · 2026-10-09
02

Host compatibility

What the documentation claims. We have not run a compatibility test.

HostStatusNotes
claude-codementioned
03

What it tells the agent

The instruction file, verbatim from the audited commit — this is the text the model reads, and the surface the audit's instruction layer examines. Quoted here so you can judge it without cloning anything.

---
name: ideogram-reference-architecture
description: >-
  Design an Ideogram service boundary with policy gateway, queue, async state, verified webhooks, polling, moderation, storage, and audit. Use when creating or reviewing a production architecture. Trigger with "architect Ideogram", "design an Ideogram service", or "review an Ideogram system diagram".
allowed-tools: Read,Glob,Grep,Write,Edit
argument-hint: "<workload> <tenant-model> <slo>"
version: 1.11.0
license: MIT
author: Jeremy Longshore <[email protected]>
tags: [saas, ideogram, architecture]
model: inherit
effort: high
compatibility: "Designed for Claude Code; architecture work is evidence-led and deployment-neutral"
---
# Ideogram Reference Architecture

## Overview

Define ownership and data flow for a production Ideogram integration. Keep user authorization, paid vendor authority, request validation, async state, event verification, content review, durable media, and audit evidence as explicit components.

## Prerequisites

- Workload, tenant and identity model, data class, SLO, budget, region, and retention policy.
- Expected endpoint mix, input sizes, concurrency, publication destinations, and failure modes.
- Owners for billing, security, moderation, storage, operations, and vendor escalation.

## Current Contract

The API uses server-side `Api-Key`, endpoint-specific multipart requests, sync and async generation, item-level safety, signed async webhooks, polling, and temporary asset URLs. Default capacity is 10 in-flight requests. These contracts require stateful application orchestration even when the product experience is synchronous.

## Authentication

Place `IDEOGRAM_API_KEY` only in a private vendor adapter. The policy gateway authenticates application identities and authorizes tenant, use case, spend, input, output, and destination separately.

## Instructions

1. Define ingress and policy gateway boundaries for identity, tenant, rights, media validation, safety, copyright settings, budget, and admission.
2. Route accepted work to a durable queue and an account-level concurrency controller.
3. Use a vendor adapter that owns endpoint schemas, multipart, deadlines, typed errors, and content-free telemetry.
4. Persist request and `generation_id` state before async delivery, then reconcile verified webhooks with polling.
5. Pass safe outputs through a downloader that validates media and stores opaque tenant-scoped objects immediately.
6. Publish only from the application object store after required review; never authorize from a vendor URL.
7. Add outbox, audit, deletion, incident, canary, and rollback paths with explicit owners.

## Tool Discipline

Use Read, Glob, and Grep to ground the design in the actual repository and infrastructure. Use Write and Edit for approved architecture records or implementation. Do not provision services, create keys, or deploy from a design invocation.

## Approval Boundaries

Require accountable review for trust-boundary changes, public ingress, data residency, sensitive media, external publication, retention, spend, and production topology. Record unresolved assumptions as blockers.

## Error Handling

- Keep queue acceptance distinct from vendor acceptance and durable asset completion.
- Reconcile duplicate or missing events without duplicate paid submissions.
- Quarantine unsafe, malformed, oversized, or tenant-mismatched media before publication.

## Output

Return components, flows, trust boundaries, state model, endpoint choices, SLO and capacity model, data lifecycle, controls, owners, failure paths, open decisions, and rollback architecture. Exclude secret or content examples.

## Examples

- Flow: application client -> policy gateway -> queue -> vendor adapter -> state store -> verified webhook or poller -> safe downloader -> object store -> reviewer -> publisher.
- Keep audit records content-free while placing regulated media in its approved retention boundary.

## Validation

Walk happy, unsafe, throttled, duplicate, missing-webhook, expired-URL, storage-failure, and rollback scenarios. Verify every state and retained datum has one owner and deletion path.

## Resources

- [Current first-party evidence map](references/official-docs.md) — use the dated endpoint, webhook, billing, team, and training links as the contract index for this workflow.
- Recheck the endpoint-specific page and current OpenAPI description before relying on an enum, limit, beta feature, or lifecycle claim.
- Record live observations as environment-specific evidence, not as universal vendor guarantees.
04

Trust audit

SAFEgrade B · trust 89/100 Nothing in the source contradicts what it says it does. Grade A is reserved for packages that have also passed the behavioural sandbox.

LayerWhat it checksResult
L0Provenance & inventoryPASS
L1Static analysis of the codePASS
L2Instruction surface (what it tells the agent)PASS
L3Class-specific surfacePASS
L4Behavioural (sandbox)SKIPPED

What the source does

Filesystem
none-observed
Network
none-observed
Shell
none-observed
Dependencies
pinned
Secrets in source
none-found

Findings (0)

No findings outside the package's declared scope.

Gates applied: no_behavioural_pass.

Audited 2026-10-09 · audit v0.4.1 · source sha 4f83675ca38afull audit observations/trust-audit/skill/jeremylongshore__ideogram-reference-architecture.json · Report an issue / request a re-scan
05

Audit history

Every audit this skill has had.

DateSourceVerdictGradeScoreChange
2026-10-094f83675ca38aSAFEB89first audit
06

Questions

What does the Ideogram Reference Architecture skill do?

Model-agnostic agent-skills platform with a harness-free canonical layer, verified adapters, and the ccpi package manager. Explore at tonsofskills.com.

Is Ideogram Reference Architecture safe to install?

The audit found nothing in the source that contradicts what it says it does, and graded it B (89/100). Grade A is held back for packages that have also passed a sandboxed behavioural run, which is why a clean skill reads B.

What can Ideogram Reference Architecture access on my machine?

The audit observed no filesystem, network or shell use at all in its source.

Which assistants does Ideogram Reference Architecture work with?

Its documentation mentions claude-code. That is what the text claims, not a compatibility test we ran.

How current is this page?

The grade is for one exact copy of the source (4f83675ca38a), read on 2026-10-09. The repository is watched, and a new audit runs when it changes — this is the first audit.

Advertisement