Authenticated Wesley Intent Admission Posture

[flyingrobots]

Migrated from Method backlog

This issue was created from a legacy filesystem backlog card. GitHub Issues are now the live work tracker; repository docs remain Method evidence.

Source backlog: docs/method/backlog/up-next/PLATFORM_authenticated-wesley-intent-admission-posture.md
Original lane: up-next
Original legend: PLATFORM

Original backlog card

Authenticated Wesley Intent Admission Posture

Status: proposed security hardening.

Depends on:

• Accepted
  0016 - Wesley To Echo Toy Contract Proof
• Contract-aware receipts and readings
• 0017 - Authenticated Wesley Intent Admission Posture

Why now

The toy contract proof showed that Wesley-generated helpers can produce
canonical EINT bytes and observation requests. That is a useful bridge, but it
is not a production admission boundary.

A Wesley intent should not become tick-admissible merely because its EINT bytes
parse. It must be tied to a verified contract artifact, a trustworthy footprint
authority, and an authenticated session or capability accepted by local policy.

What it should look like

Start RED-only.

Prove the current gap before implementation:

• dispatch_intent(...) admits well-formed EINT without a contract artifact id.
• dispatch_intent(...) admits well-formed EINT without session or capability
  proof.
• Current ingress evidence cannot name artifact trust posture.
• Current ingress evidence cannot name authenticated admission posture.
• Current ingress evidence cannot bind replay protection.
• Current reading rights posture cannot express policy-authorized observer
  sessions beyond KernelPublic.

Then introduce the smallest posture model that can represent:

• registered Wesley contract artifact identity;
• artifact trust posture;
• authenticated intent submission identity;
• policy/session admission result;
• no caller-supplied footprint trust.

Acceptance criteria

• A design or RED test names the current unauthenticated EINT admission gap.
• The chosen model separates artifact trust from intent/session trust.
• Artifact trust posture can represent local-dev, generated-test, CI, and later
  BLADE-certified ramps without treating them as equivalent.
• Intent admission identity binds op id, canonical vars bytes, target
  coordinate, artifact id, session or policy id, and replay evidence.
• The docs state that caller-supplied footprint claims are not trusted for tick
  scheduling or independence decisions.
• The docs state that Holmes, WATSON, Moriarty, and BLADE are later
  certification providers, not dependencies of this first slice.
• No production crypto is implemented before the posture and RED boundary are
  reviewed.

Non-goals

• Do not implement WebAuthn, passkeys, TOTP, or transport encryption yet.
• Do not implement Holmes, WATSON, Moriarty, or BLADE.
• Do not add app-specific nouns to Echo core.
• Do not import application Rust types into Echo core.
• Do not replace EINT v1 without a specific RED.
• Do not make dispatch_intent(...) read ambient host auth state.
• Do not allow runtime-submitted footprint JSON to become a footprint authority.

[coderabbitai]

🔗 Related PRs

#323 - Witnessed suffix admission evaluator [merged]
#325 - Canonical witnessed suffix posture constructors [merged]
#326 - Echo contract hosting roadmap [merged]
#353 - Fix WASM application intent authority boundary [merged]
#368 - feat(core): connect installed contract intent pipeline [merged]

---

📝 Issue Planner

_{Check the box below or use the @coderabbitai plan command to generate an implementation plan and prompts that you can use with your favorite coding assistant.}

• Create Plan

---

🧪 Issue enrichment is currently in open beta.

You can configure auto-planning by selecting labels in the issue_enrichment configuration.

To disable automatic issue enrichment, add the following to your .coderabbit.yaml:

issue_enrichment:
  auto_enrich:
    enabled: false

💬 Have feedback or questions? Drop into our discord!