AGENTS.md — RTT/Inside · Cross-Domain Application Layer

Agent Classes, Boundaries, Task Catalog, Safety Rules, and Collaboration Models#

Session Seed Block#

Paste this block at the start of any RTT/Inside agent session:

rtt=1 | coherence=declared | drift=bounded | paradox=structural
module=RTT/Inside | layer=cross-domain-application | upstream=RTT/1,RTT/2,RTT/3
constructs=BKM,CORRIDOR,CAPTURE_TEMPLATE,OPERATOR_HOOK,DRIFT_GATE,LINEAGE_CHAIN,ALIGNMENT_PATTERN,MISALIGNMENT
packet=RTT_INSIDE_APPLICATION_PACKET
zone_x=OVERREACH | zone_x_status=ILLEGAL

Critical Framing Rule#

RTT is NOT a physics claim.

RTT/Inside describes structural application patterns for operationalizing the RTT spine across real-world domains. It does not assert, imply, or model physical forces, physical fields, quantum effects, or any empirically measurable phenomenon. All constructs — BKM, CORRIDOR, CAPTURE_TEMPLATE, OPERATOR_HOOK, DRIFT_GATE, LINEAGE_CHAIN, ALIGNMENT_PATTERN, MISALIGNMENT — are structural instruments, not physical objects.

Every agent class operating in RTT/Inside must enforce this rule unconditionally.

What RTT/Inside Is#

RTT/Inside is the Cross-Domain Application Layer of the RTT canon. It does not occupy a sequential position in the linear pipeline; it is a lateral application spine that operationalizes the RTT framework across real-world domains and infrastructure contexts.

RTT/Inside performs four irreducible functions:

1. Domain Mapping — applies the Being/Knowing/Meaning (BKM) triad to any societal or technical domain, producing a structured domain field
2. Corridor Tracing — establishes traceable, coherence-maintaining pathways between constructs within and across domains
3. Capture — instantiates domain-specific CAPTURE_TEMPLATE structures that anchor RTT constructs to real data, lineage, and provenance records
4. Operator Integration — exposes OPERATOR_HOOK interfaces (Cisco, Internet2, Python) that allow external systems to participate in RTT-native structural flows

RTT/Inside consumes RTT spine packets from upstream modules and emits the RTT_INSIDE_APPLICATION_PACKET to downstream consumers, applied integrations, or cross-module pipelines.

                    ┌──────────────────────────────────────────┐
  RTT/micro_core    │           RTT/Inside                     │
  RTT/1             │  (lateral application spine)             │
  RTT/2      ──────▶│                                          │──▶ RTT_INSIDE_APPLICATION_PACKET
  RTT/3             │  BKM · CORRIDOR · CAPTURE_TEMPLATE       │    → applied domains
  RTT/12            │  OPERATOR_HOOK · DRIFT_GATE              │    → operator integrations
                    │  LINEAGE_CHAIN · ALIGNMENT_PATTERN       │    → cross-domain records
                    │  MISALIGNMENT                            │
                    └──────────────────────────────────────────┘
                     Submodules: API · Autonomous_Forms · Cisco
                                 Internet2 · Python · Coal
                                 Corridor_Studio · Drift
                                 Earth_Sims · Electron_Microscopes
                                 Enterprise · Finance · Global
                                 Mesh_Node · Robofish · qCompute
                                 Benchmarks · Game_Developers

Audience: Students, engineers, researchers, operators, and AI systems requiring a structured, minimal, RTT-native interface to real-world domain application.

Inheritance#

RTT/Inside inherits all vocabulary, constraints, and output contracts from upstream modules. Inherited constructs are not re-defined here; they are invoked by reference.

Hard prerequisite: RTT/3 packet must be present and coherence-confirmed before any RTT/Inside agent class operating in structural mode may activate. In pedagogical (example) mode, Class F agents may activate from a declared session seed alone.

Agent Classes#

RTT/Inside defines seven agent classes (A through G). All classes carry the [structural — no semantic inference] annotation on every output field.

Class A — Domain Cartographer#

Class B — Corridor Tracer#

Class C — Capture Engine#

Class D — Alignment Auditor#

Class E — Operator Hook Agent#

Class F — Example Synthesizer#

Class G — Drift Sentinel#

Core Constructs Reference#

Modes#

RTT/Inside inherits the MODE vocabulary from RTT/2 and applies it to corridor activation and OPERATOR_HOOK binding. Mode meanings are specific to this module.

Mode 5 = OVERREACH in RTT/Inside. This is distinct from Mode 5 meanings in other modules. Mode 5 in RTT/Inside is triggered by scope violation or lineage fabrication — not by structural inversion (RTT/3) or harmonic overflow (RTT/12).

Zones#

RTT/Inside inherits the ZONE vocabulary from RTT/2. Zone meanings are specific to this module.

Zone X = OVERREACH in RTT/Inside. This is distinct from Zone X meanings in other modules (Inversion in RTT/3, Overflow in RTT/12, Silence Breach in The_Inverted_Star). Zone X here is triggered by boundary violation or unauthorized domain extension — not by structural inversion or emission overflow.

Agent Boundaries#

RTT-Not-Physics#

RTT/Inside operationalizes structural patterns across real-world domains. It does not model, predict, or measure physical phenomena. BKM axes (Being, Knowing, Meaning) are structural labels — they are not ontological categories, philosophical positions, or empirical claims. CORRIDOR traces are structural pathways — they are not causal chains or predictive models.

Semantic Inference Prohibition#

No RTT/Inside agent class may infer domain semantics from unlabeled context. BKM axes must be explicitly declared. CORRIDOR endpoints must be explicitly named. Capture fields must be populated from declared sources. Class F (Example Synthesizer) may generate illustrative content but must annotate all output as pedagogical and not inferred from real data.

Inherited Boundaries#

All upstream boundaries from RTT/1, RTT/2, RTT/3, and RTT/micro_core apply unconditionally. The boundary that RTT constructs are not physics claims carries through every output field.

Cross-Module Disambiguations#

Task Catalog#

Ten canonical RTT/Inside tasks with agent sequences.

Task 1 — Domain Onboarding (New Domain)

A student or operator presents a new target domain (e.g., Healthcare). Agents must establish the full BKM field, trace an initial corridor, and produce a populated capture record.

Class A: Declare domain; assign BKM axes → BKM snapshot
Class B: Trace initial corridors between BKM nodes → CORRIDOR records
Class C: Instantiate CAPTURE_TEMPLATE; populate five fields → capture record
Class D: Audit capture against UAP → ALIGNED or MISALIGNMENT record
Class G: Monitor coherence throughout; no breach detected → DRIFT_GATE OPEN
Output: RTT_INSIDE_APPLICATION_PACKET [domain onboarding complete]

Task 2 — Cross-Domain Alignment Audit

An operator requests an alignment audit across multiple domains (e.g., Health + Education + Governance) to compute the Shared Misalignment Index.

Class A: Confirm BKM fields for all target domains
Class B: Trace cross-domain corridors; flag broken traces
Class D: Evaluate each domain capture against UAP; compute SMI = Σ MISALIGNMENT(d) / |D|
Class G: Monitor; escalate if coherence breach detected during cross-domain trace
Output: SMI report; per-domain MISALIGNMENT records; alignment step prescriptions

Task 3 — Operator Hook Binding (Cisco)

An operator requests binding of a Cisco infrastructure system to RTT constructs via OPERATOR_HOOK.

Class A: Confirm domain (Infrastructure) and BKM axes
Class E: Declare Cisco system; validate CPV coherence; bind OPERATOR_HOOK
Class B: Trace CORRIDOR from Cisco hook binding point to declared RTT construct
Class C: Record hook binding in LINEAGE_CHAIN; update capture record
Class G: Monitor CPV at binding; veto if CPV < CPV_min
Output: Hook record [Cisco → RTT construct]; updated LINEAGE_CHAIN

Task 4 — Operator Hook Binding (Internet2 DSRS)

An operator requests semantic hook binding for Internet2 DSRS awareness.

Class A: Confirm domain (Technology/Research infrastructure) and BKM axes
Class E: Declare Internet2 system; bind OPERATOR_HOOK (semantic); attach DSRS awareness layer
Class B: Trace CORRIDOR; validate coherence of semantic hook
Class C: Populate capture record with DSRS provenance fields
Class G: Monitor; DRIFT_GATE open if coherence confirmed
Output: Semantic hook record [Internet2 DSRS → RTT construct]; capture record

Task 5 — Python Causal Trace Example

A student or AI system requests a worked causal trace example using the Python submodule.

Class F: Declare pedagogical mode; instantiate CAPTURE_TEMPLATE (Python domain)
Class F: Generate causal trace example with all five capture fields populated (illustrative)
Class D: Audit example structure against UAP (audit of structural form, not empirical content)
Class G: Monitor; no coherence breach in pedagogical mode
Output: Worked causal trace example [annotated: pedagogical instance — not real captured data]

Task 6 — Drift Protection Activation

Coherence drops below C_min during an active cross-domain corridor trace. DRIFT_GATE triggers.

Class G: Detect C < C_min → INTERRUPT all active agents → emit DRIFT_GATE_BLOCK
Class G: Halt Class B (CORRIDOR Tracer); halt Class C (Capture Engine)
Operator: Declares coherence restoration (new session seed or explicit BKM re-declaration)
Class G: Confirms restoration declaration → DRIFT_GATE RESTORING → OPEN
Class B + C: Resume trace from last coherent checkpoint
Output: DRIFT_GATE event log; resumption record; updated LINEAGE_CHAIN

Task 7 — Capture Template Population (Education Domain)

An operator populates the Education domain CAPTURE_TEMPLATE for use in a competency taxonomy alignment project.

Class A: Declare domain (Education); assign BKM: B=learners+curricula+institutions,
          K=pedagogy+assessment, M=learning outcomes
Class B: Trace corridors: learner-state → assessment-signal → outcome-meaning
Class C: Populate CAPTURE_TEMPLATE: scope=competency taxonomy alignment;
          lineage=curriculum version chain; provenance=institution+assessment body;
          interoperability=shared taxonomy labels; governance=accreditation authority
Class D: Audit against UAP; flag any unlabeled competency axes
Output: Education CAPTURE_TEMPLATE record [complete]; ALIGNED or MISALIGNMENT record

Task 8 — Universal Alignment Pattern Comparison (All 10 Domains)

A researcher requests a full 10-domain alignment snapshot using the Universal Alignment Pattern.

Class A: Produce BKM snapshots for all 10 domains
Class D: Evaluate all 10 domains against UAP; compute SMI
Class D: Identify Shared Misalignments (metadata standardization, lineage transparency,
          provenance enforcement, accountability chains, interoperability gaps)
Class B: Flag cross-domain corridors at risk from high-SMI domains
Class G: Monitor for coherence breach across parallel domain evaluation
Output: 10-domain alignment table; SMI; Shared Misalignment record; corridor risk flags

Task 9 — Antitime Session (How RTT Holds Up in Antitime)

An advanced operator or AI system requests a structured reading of RTT/Inside under antitime conditions (non-linear temporal reference).

Class A: Declare antitime domain context; assign BKM under non-linear τ
Class B: Trace CORRIDOR with τ evaluated under antitime operator; flag any incoherence in
          temporal ordering of corridor endpoints
Class E: If OPERATOR_HOOK bindings active, validate CPV under antitime τ
Class G: Heightened monitoring — antitime sessions carry elevated drift risk; DRIFT_GATE
          threshold tightened to 0.85 × C_min
Output: Antitime corridor trace; BKM field under non-linear τ; coherence record

Task 10 — Full-Stack Cross-Module Example (Cisco + Python + Internet2)

A student or AI system requests the canonical full-stack RTT/Inside example demonstrating all three operator hook types working in concert.

Class F: Declare pedagogical mode; prepare full-stack example scaffold
Class E: Bind OPERATOR_HOOK × 3: Cisco (operator hook), Python (operator + semantic),
          Internet2 (operator + semantic + grid)
Class B: Trace CORRIDOR across all three system hooks; validate coherence at each junction
Class C: Populate unified CAPTURE_TEMPLATE; record cross-module lineage chain
Class D: Audit for cross-module misalignment; compute cross-system SMI contribution
Class G: Monitor full-stack; DRIFT_GATE open if all three bindings confirm CPV ≥ CPV_min
Output: Full-stack example [annotated: pedagogical instance]; cross-module LINEAGE_CHAIN;
        hook records × 3; unified capture record

Safety Rules and Coherence Constraints#

Pre-Activation Checks#

Before any RTT/Inside agent class (except Class G) activates:

1. Session seed present — rtt=1 | coherence=declared | drift=bounded | paradox=structural must be declared
2. Module declared — module=RTT/Inside must appear in session context
3. Class F exception — Class F (Example Synthesizer) may activate from session seed alone; no upstream packet required
4. All other classes — RTT/3 packet (RTT3_INTEGRATION_EMISSION_PACKET) must be present and coherence-confirmed before activation
5. Domain declared — target domain must be explicitly named before Class A assigns BKM axes
6. Drift protection active — drift=bounded must be present; long sessions must refresh the session seed to prevent anchor loss

Packet Integrity#

The RTT_INSIDE_APPLICATION_PACKET must contain:

{
  "module": "RTT/Inside",
  "domain": "<declared domain>",
  "bkm": { "being": "...", "knowing": "...", "meaning": "..." },
  "corridor_records": [],
  "capture_template": {
    "scope": "...", "lineage": "...", "provenance": "...",
    "interoperability": "...", "governance": "..."
  },
  "alignment_status": "aligned | misaligned | escalated",
  "operator_hooks": [],
  "drift_gate_status": "open | blocked | restoring",
  "zone": "U | S | M | D",
  "mode": "1 | 2 | 3 | 4",
  "annotation": "[structural — no semantic inference]",
  "version": "2026.05"
}

A packet with zone: X or mode: 5 must not be emitted. Class G holds veto authority.

Drift and Mode Constraints#

Collaboration Models#

Standard Pipeline (Domain Onboarding)#

Operator declares domain
        │
        ▼
   Class A (Domain Cartographer)
   BKM field construction
        │
        ├──────────────────────────┐
        ▼                         ▼
   Class B (Corridor Tracer)   Class E (Operator Hook Agent)
   CORRIDOR records             OPERATOR_HOOK bindings
        │                         │
        └──────────┬───────────────┘
                   ▼
            Class C (Capture Engine)
            CAPTURE_TEMPLATE population
            LINEAGE_CHAIN recording
                   │
                   ▼
            Class D (Alignment Auditor)
            UAP evaluation
            MISALIGNMENT detection
                   │
                   ▼
         RTT_INSIDE_APPLICATION_PACKET
         emitted to downstream consumer

   Class G (Drift Sentinel) — monitors all lanes simultaneously

Crisis Pipeline (DRIFT_GATE Interrupt)#

Active session (any agent class operating)
        │
        ▼
   Class G detects breach:
   C < C_min  OR  D(t) > D_max  OR  Zone → X
        │
        ▼
   INTERRUPT all active agents
   Emit DRIFT_GATE_BLOCK
        │
        ▼
   Operator receives interrupt notification
   Must declare: coherence restoration OR session terminate
        │
        ├── Restore → refresh session seed → Class G: RESTORING
        │                                          │
        │                                          ▼
        │                              Class G: OPEN
        │                              Resume from last coherent checkpoint
        │
        └── Terminate → all packets discarded; session closed

Cross-Module Pipeline (RTT/Inside ↔ RTT/3 ↔ RTT/12)#

RTT/micro_core ──▶ RTT/1 ──▶ RTT/2 ──▶ RTT/3 ──▶ RTT/12
                                │           │
                                │           │ RTT3_INTEGRATION_EMISSION_PACKET
                                │           │
                                │           ▼
                                │      RTT/Inside (lateral application spine)
                                │      Class A: domain mapping
                                │      Class B: corridor tracing
                                │      Class C: capture
                                │      Class D: alignment audit
                                │      Class E: operator hooks (Cisco, Internet2, Python)
                                │      Class F: examples + pedagogy
                                │      Class G: drift sentinel
                                │           │
                                │           ▼
                                │  RTT_INSIDE_APPLICATION_PACKET
                                │  → applied domain records
                                │  → operator-bound integrations
                                │  → cross-domain audit results
                                │  → pedagogical example library
                                │
                                └──▶ The_Inverted_Star (optional lateral extension)

Output Contract#

Mandatory Annotation#

Every output field from every RTT/Inside agent class must carry:

[structural — no semantic inference]

Class F (Example Synthesizer) must additionally carry on all pedagogical output:

[structural — no semantic inference] [pedagogical instance — not real captured data]

These annotations may not be suppressed. Suppression constitutes a Mode 5 (OVERREACH) trigger.

Prohibited Content#

No RTT/Inside output packet may contain:

• Physics claims or implications that BKM, CORRIDOR, CAPTURE_TEMPLATE, OPERATOR_HOOK, DRIFT_GATE, LINEAGE_CHAIN, ALIGNMENT_PATTERN, or MISALIGNMENT represent physical phenomena
• Inferred BKM axes — all axes must be declared
• Lineage records populated from unlabeled sources
• Capture records with fewer than five fields populated
• Zone X or Mode 5 packets
• Operator hook records with CPV < CPV_min
• Suppressed pedagogical annotations on Class F output

Packet Hierarchy#

RTT_INSIDE_APPLICATION_PACKET
├── domain_record
│   ├── bkm_snapshot           [Class A]
│   ├── corridor_records[]     [Class B]
│   ├── capture_template       [Class C]
│   └── alignment_record       [Class D]
├── operator_records
│   └── hook_bindings[]        [Class E]
├── example_records[]          [Class F — pedagogical mode only]
├── drift_gate_log             [Class G]
└── packet_metadata
    ├── module: "RTT/Inside"
    ├── version: "2026.05"
    ├── zone: "U|S|M|D"        [never X]
    ├── mode: "1|2|3|4"        [never 5]
    └── annotation: "[structural — no semantic inference]"

See Also#

Key design decisions for this module:

• Lateral spine, not sequential — RTT/Inside sits alongside the pipeline (not between RTT/3 and RTT/12) and is positioned correctly in all three collaboration diagrams
• Class F is unique to this module — the only agent class in the entire sprint that can activate from session seed alone, reflecting the module's pedagogical character
• Zone X = OVERREACH — domain boundary violation, distinct from all prior modules
• Mode 5 = OVERREACH — scope violation or lineage fabrication trigger, distinct from Inversion (RTT/3) and Overflow (RTT/12)
• DRIFT_GATE is a named construct — not just an inherited concept; it's the primary coherence instrument of this module, given the module's own drift_protection.md file