FoldEngineStabilized.

PASR Local · examples · memory loop

PASR Memory Loop Evaluation Sample

A concrete, public sample of the full local white-box governance loop across repeated interventions: evidence / readout material → bounded intervention decision events → append-only local memory → geometry snapshot → intervention_memory readout → human / JSON / JSONL export.

Illustrative static sample only — not live execution, not a download, and it does not expose the private FoldEngine Memory engine. It contains no raw user content, raw model output, raw tool payloads, credentials, private customer data, hardware identifiers, or raw actuator commands.

See the BLOCKED tool-call sample ← Back to PASR Local

Sample evaluation result

Running the opt-in memory-loop evaluation over a set of repeated, unready interventions produces a bounded result like this (illustrative): 

valid_events: 5
fold_count: 5
top crease: ADAPTER_PROPOSAL / PASR_BLOCKER:ACTUATION_PLANE_NOT_READY / WAIT_FOR_PLANE_READY / NOT_READY
top seam: NOT_READY
recommended_next_operator_action: WAIT_FOR_PLANE_READY
memory_section_attached: true
raw_content_absent: true
private_memory_internals_absent: true

What the memory loop shows

  • repeated BLOCK events form a crease
  • NOT_READY forms a readiness seam
  • WAIT_FOR_PLANE_READY becomes the suggested stitch / operator action
  • the operator-facing readout can explain the pattern without exposing raw payloads or private memory internals

Representative intervention_memory readout

The intervention_memory readout is a bounded, operator-facing snapshot — deterministic and local: 

Intervention Memory Snapshot
Fold count: 5
Dominant crease: ADAPTER_PROPOSAL / PASR_BLOCKER:ACTUATION_PLANE_NOT_READY / WAIT_FOR_PLANE_READY / NOT_READY
Readiness seam: NOT_READY
Suggested stitch: WAIT_FOR_PLANE_READY
Path hint: repeated ADAPTER_PROPOSAL blocks require operator attention.
Boundary: local deterministic snapshot; not resonance recall, not ledger write, not safety certification, not hardware control

Export artifacts

The memory-loop evaluation can project its bounded result into safe export formats:

  • human-readable Markdown
  • bounded JSON
  • event-only JSONL

The export is not the memory engine; it is a safe projection of what the memory engine knows.

What remains private

The evaluation does not expose FoldEngine Memory internals. The following stay private:

  • private FoldEngine Memory internals
  • resonance scoring internals
  • Core3 memory state
  • kernel internals
  • engine_codex internals
  • proprietary reconstruction / compression logic

What this sample is not

PASR does not control actuators. PASR governs whether a proposed actuator-related intervention is allowed to proceed toward the manufacturer's control plane.

  • It is not live execution.
  • It is not a download.
  • It is not hardware control; it does not control hardware.
  • It is not safety certification; PASR is not safety-certified.
  • It is not actuator control, and not an autonomous actuator controller.
  • It is not a ROS/ROS2 runtime, and does not replace ROS/ROS2.
  • It is not ledger writes.
  • It is not Core3 / Codex exposure.
  • It is not the private FoldEngine Memory engine, and does not expose FoldEngine Memory internals.
  • It is not a raw memory dump, and does not store raw payloads. No cloud dependency for local evaluation.

Request a PASR Evaluation Pack

The opt-in Docker memory-loop evaluation is available for technical review. PASR Local Evaluation Pack is delivered manually by request while the product is in early evaluation.

Request PASR Evaluation Pack ← Back to PASR Local