Fitrah

Inspiration

Old ideas about trust, tested against computer science

Fitrah takes its name, and several research prompts, from Islamic thought about trust, stewardship and accountability. This page shows how each idea maps onto models computer science already has. Updated 2 Nov 2026

A rule we hold ourselves to: naming a familiar access-control technique after a beautiful concept doesn’t make it new. Every inspiration is checked against prior art first. These are human moral and legal teachings; none of them is a specification for software, and nothing here is a religious ruling.

The mapping

Each idea, and what computer science already knows

ConceptThe ideaClosest computer-science modelsWhat it means for Fitrah
فطرةFitrahInnate nature; the disposition one is born withMandatory access control; policy reference monitors (Cedar, OPA)The inherited core. Not new on its own.
أمانةAmanahA trust held on behalf of others, with accountabilityAttribute-based access control, usage control (UCON), capability systemsThe split between authority an agent can use and authority it may pass on. Responsibility here is not moral personhood.
وقفWaqfAn endowment whose conditions outlast the people who administer itUCON and XACML obligations, decentralised information-flow labelsA candidate question: obligations that survive an agent’s restart or replacement (below).
إجازةIjazahLicence to teach, including permission to license othersSPKI/SDSI, macaroons, Biscuit, XACML delegation profileRestricted delegation chains already exist. Not unique to authorised birth.
إسنادIsnadA chain of transmitters behind a reportW3C PROV, Verifiable Credentials, in-toto, SLSAA signed chain shows who said what, not that it’s true.
حسبةHisbahIndependent public scrutiny of conductReference monitors, remote attestation verifiers (RFC 9334)Keep enforcer and verifier separate. Already established practice.
شورىShuraConsultation before consequential decisionsSeparation of duty, multi-party approvalA possible control for amending inherited rules. Not part of Lab 0.
معروفMa’rufRecognised good that takes precedenceDeny-overrides policy combination, mandatory access controlHigher limits override requests. Already reflected in Fitrah’s explicit allow rule.

Scale isn’t novelty either. Google’s Zanzibar already checks relationship-based permissions at enormous scale, and agent delegation is an active research area (IETF WIMSE and 2026 preprints). The bar for a new contribution is high.

A candidate research question

Obligations that outlive the agent

A waqf separates the people who administer an endowment from the conditions that govern it. That suggests a question for agents: when an agent restarts, is replaced or delegates, should the obligations attached to a resource survive, and can that be enforced?

The question

  • Under crashes, retries, delegation, credential rotation and revocation, can resource obligations be preserved and enforced?
  • Example: a healthcare agent processes records for a declared purpose and delegates to a child that may restart. The purpose restriction must survive.

How it would be tested fairly

  • Compare a strong usage-control broker with obligations against the same broker plus Fitrah’s hereditary rules, on identical fault traces.
  • If the strong broker achieves the same properties, report equivalence instead of renaming the same rules.
  • Only checkable resource operations: no claims about ethics or intent.

Interpretive limits. Waqf practice developed over centuries, with recognised and disputed adjustments, so “never amend” is not an accurate reading. Usage-control obligations also predate Fitrah. Whether this question leads anywhere new is open.

Help wanted

We need scholars as much as engineers

This line of work needs review from a qualified scholar of Islamic jurisprudence for the sources, and a historian of waqf for the institutional history, alongside security researchers to challenge the computer-science side.

Get in touch See the findings

Sources on trust and stewardship