# PrismManifest One canonical identity. Human page: https://www.insightits.com/products/prismmanifest.html | Field | Value | |-------|-------| | Slug | `prismmanifest` | | Status | published | | Family | aiSecurity | | Infra category | AI Security | | Record type | product | | Parent | — | | JSON | https://www.insightits.com/catalog/prismmanifest.json | | Markdown | https://www.insightits.com/catalog/prismmanifest.md | | GitHub | https://github.com/insightitsGit/PrismManifest | | PyPI | https://pypi.org/project/prismmanifest/0.3.4/ | | Install | `pip install prismmanifest (pin ==0.3.4). Apache-2.0. Soft CTA MANIFEST — mailto:info@insightits.com?subject=MANIFEST.` | | Version | 0.3.4 | ## What it is Zero-trust tool-argument gate. AI/OCR may propose dollar parameters; only an Ed25519-signed ParameterManifest that clears the hard security check may enter tax, underwriting, claims, or payment engines. Formerly ParamGate. Works with any orchestrator — ChorusGraph optional. ## Problem Deterministic Engine + Unverified Probabilistic Input = Deterministic Wrong Answer. Fluent OCR/LLM dollars must not execute in money engines. ## Alternatives - **Ungated LLM/OCR → engine** — The alternative in production is treating extracted dollars as trusted input. PrismManifest is the signed-manifest boundary; there is no published competitor bake-off page. ## Features - **Tool-Argument Gate** — Sits between LLMs/OCR and your tax, underwriting, claims, or payment engine so unverified money never enters the run. - **Signed ParameterManifest** — Ed25519-signed manifests only — not “0.87 confidence” as permission to pay. - **Fail-Closed Boundary** — Hard security check (C++ and/or Python twin) verifies the signed manifest before the engine runs. Soft CTA: MANIFEST. ## Architecture Evidence spans → allow / human / refuse → Ed25519 ParameterManifest → hard check (C++ and/or Python twin) before the engine. Integrity stack: PrismGuard (who may speak) · Prism-Eval (CI) · PrismManifest (which numbers may execute) · Prism-Shield (runtime gateway) · PrismShine (grounding). ## Use cases - Tax engines (Form 1040 Line 11 AGI lab) - Underwriting, claims, and AP money arguments (pattern tabs are illustrative) - Any orchestrator that must not pass unverified dollars into a deterministic engine ## Benchmarks Fixed sample Form 1040 AGI $450,000 fixture used by the Digit Drop Lab and interactive demo. ### Disclosures - Vendor-authored evidence dataset. Do not invent additional metrics. ### Evidence URLs - https://www.insightits.com/products/digit-drop-lab.html ## GitHub https://github.com/insightitsGit/PrismManifest ## PyPI https://pypi.org/project/prismmanifest/0.3.4/ ## Documentation - [Digit Drop Lab](https://www.insightits.com/products/digit-drop-lab.html) - [Money Path Demo](https://www.insightits.com/products/prismmanifest-demo.html) - [Compose with VectorPrism / Shield / Eval](https://www.insightits.com/products/prismmanifest.html#compose) - [https://github.com/insightitsGit/PrismManifest](https://github.com/insightitsGit/PrismManifest) - [https://pypi.org/project/prismmanifest/0.3.4/](https://pypi.org/project/prismmanifest/0.3.4/) - [https://www.insightits.com/products/prismguard.html](https://www.insightits.com/products/prismguard.html) - [https://www.insightits.com/products/prismshine.html](https://www.insightits.com/products/prismshine.html) ## Is not - An LLM - Full OCR - PrismGuard (who may speak) - PrismShine (whether the answer is grounded) - An agent runtime ## Related products - [digit-drop-lab](./digit-drop-lab.md) - [prismmanifest-demo](./prismmanifest-demo.md) - [prism-eval](./prism-eval.md) - [prism-shield](./prism-shield.md) - [vectorprism-demo](./vectorprism-demo.md)