Skip to main content
Work index
partial-implementationWorking demonstrationMixedCS-W-001

Agentic Accessibility

Parseable is not the same as legible.

Publicly bounded evidence. The page distinguishes an artifact, prototype, specification, or test from broader claims that remain under research.

Identifier
CS-W-001
Type
Working demonstration
Status
partial-implementation
Published
Updated
Author
Cognitive Studio
Working field · OARSH-CT lineagepartial-implementation
Agent-facingEnvironment
  1. OOrientable
  2. AAddressable
  3. RRelational
  4. SStateful
  5. HHabitable
  6. CComposable
  7. TTraceable
A working environmental model for orientation, authority, currency, re-entry, composition, and traceable effects. The categories remain under evaluation.

Agents can receive pages, tools, files, and structured data while still misidentifying authority, treating proposed or superseded state as current, or requiring a human to reconstruct the environment before safe action is possible.

Environmental structure should reduce reconstruction burden without creating a more dangerous source of confident error. Authority and currency must be legible, and every machine projection must remain faithful to the human-readable source it represents.

A working Agentic Accessibility model; the evolving OARSH-C → OARSH-CT design lineage; a developer guide; machine-readable coverage disclosure; semantic regions and structured entities; a tested adjacent authorization proof in The Authorized Path; and this site as a partial static working demonstration.

What this work does not establish.

  • 01

    No public cross-model benchmark yet shows that the proposed environmental mechanisms reduce agent error or human reconstruction burden.

  • 02

    The OARSH-C / OARSH-CT vocabulary is an evolving design lineage, not a validated ontology, conformance standard, or certification scheme.

  • 03

    Habitable, Composable, and Traceable remain properties that require operational tests; their presence cannot be inferred from labels, markup, or a self-assessed coverage file.

  • 04

    The current site demonstrates only the subset applicable to a static publication environment; independent human-accessibility and agent-experience review remain ahead.

The important question is not whether an agent can parse the environment. It is whether a cold agent can tell what is current, what governs, what it may do, and when the environment itself is wrong.

What the public record currently supports.

These signals are intentionally bounded. They show that work exists without inflating maturity or exposing protected source context. Where no public resource or evidence packet is linked, treat the statement as a Studio claim awaiting public source publication.

  1. 01

    The public site exposes named semantic regions, structured entities, discovery metadata, explicit work status, claim boundaries, and a machine-readable coverage record.

  2. 02

    The developer guide separates environmental properties, implementation requirements, projection integrity, security boundaries, validation layers, and known gaps rather than treating a property list as conformance.

  3. 03

    The Authorized Path provides a bounded tested result for the action boundary: authorization is checked where effects occur, bypass attempts fail, and evidence remains reconstructable.

  4. 04

    Independent agent review converged on authority legibility, currency, reconstruction burden, and projection integrity as the strongest testable directions; that review is critique, not validation.

Publish a pre-registered cold-start pilot for authority and currency, measure reconstruction burden across model families, and ship a projection-integrity checker that compares visible pages with JSON-LD, llms.txt, and other machine projections.

Related research · Authority and Action (CS-RC-001) · Agentic Accessibility (CS-RC-003) · Continuity and Return (CS-RC-004)

Continue

CS-W-002

Situated Memory