Skip to harness content
Open technical reference map

Controlled terminology

Glossary and claim boundaries

Each term has one public meaning, a status, and a usage boundary. Status identifies whether the term is standard, emerging, vendor-specific, internal, contested, or uncertain.

Term statuses

standardemergingvendor-specificinternalcontesteduncertain

Controlled terms

repository-level SDLC assurance harness for AI coding agents

emerging

Software that installs repository controls, routes work, records external coding-agent activity as evidence, checks required proof, and blocks unsupported completion.

primary_public_category

repository-level policy and evidence control plane around external coding agents

emerging

The part that coordinates policy, run state, evidence, gates, and approvals around a defined repository workflow.

qualified_secondary_architectural_role

connector-first

internal

Integration design that prioritizes normalized runtime events and artifacts.

internal_design_only

proof-first

internal

Operating principle that requires evidence before completion.

operating_principle

coding agent

emerging

External tool that can inspect a repository and perform implementation work.

external_runtime_role

Layer 0

internal

Unnumbered foundation assurance that precedes production-domain evaluation.

foundation_assurance

13 production domains

internal

Production-readiness assurance categories derived from the attributed source model.

assurance_taxonomy

Proof levels

01

Structural

Required artifacts, schemas, bindings, and coverage are valid.

02

Semantic

Commissioned adapters and thresholds prove the intended behavior.

03

Authoritative

An independent trusted runner, provider, signer, or authority attests the result with rollback-resistant state.