# llms-full.txt - Extended AI Language Model Context for CMD+RVL
# Last Updated: 2026-08-05
# https://cmdrvl.com/llms-full.txt
#
# This file provides comprehensive context for AI agents that need detailed information.
# For a concise overview, see /llms.txt

================================================================================
COMPANY OVERVIEW
================================================================================

CMD+RVL helps teams ship working AI outcomes on live business problems. Cairn is
the front door for raw questions and source material. Discovery qualifies clear
outcomes for a focused workshop. The shipped result keeps proof back to source
through metadata, lineage, and evidence records.

Website: https://cmdrvl.com
Contact: https://cmdrvl.com/contact/
Industry: Working AI Outcomes, Data Intelligence, Metadata, and Provenance

Core Value Proposition:
- CMD+RVL ships working pages, workflows, monitors, data products, and custom apps.
- Discovery starts with a 15-minute qualification call and moves to a two-hour outcome workshop when there is a fit.
- Every shipped result carries source context, metadata, lineage, and proof.
- Cairn shapes raw threads, file piles, and vendor questions before they become discovery candidates.

================================================================================
GLOSSARY OF TERMS
================================================================================

Cairn
  CMD+RVL's email-first data intelligence agent. Cairn keeps the initial
  question, source material, metadata context, and review trail together so a
  raw ask can become a page, brief, workflow, pilot, or discovery candidate.

Working AI Outcome
  A usable result shipped against a live business problem, with enough source
  context and proof to inspect, reuse, or extend it.

Discovery
  A 15-minute qualification call. If there is a fit, the next step is a
  two-hour workshop, preferably in person, with the owner and source material
  in the room.

Reference System
  A parallel execution maintained by CMD+RVL that clients can compare against
  to detect drift in their internal implementations.

Verification Loop
  The four-phase model: Produce, Replicate, Compare, Verify. This keeps outcomes
  remain valid over time.

Drift Detection
  The process of identifying when internal results diverge from the CMD+RVL
  reference system.

Evidence Pack
  A structured artifact attached to every outcome delivery containing:
  - Source filing links and accession numbers
  - Publish/ingest timestamps
  - Prior state pointers
  - Methodology version
  - Transformation steps
  - Delivery logs

State
  The current disclosure-derived state about a monitored entity, with
  source links, timestamps, and versions.

State Transition
  A structured, deterministic event emitted when monitored state changes.
  Includes prior state, current state, delta summary, and evidence links.

Signals
  CMD+RVL's discovery surface where users explore market data, test hypotheses,
  and identify patterns before committing to formal outcomes.

DealCharts
  A public evidence surface that shows how structured finance data is
  constructed, with provenance links back to source filings.

Disclosure-Derived
  Data or state that is constructed exclusively from publicly filed disclosures.
  No proprietary data sources and no inference. Only disclosed facts.

Admissibility
  Whether evidence meets the trust threshold required for a given context.
  Varies by use case: regulatory, audit, operational.

Agent-Ready
  Outcomes and transitions designed for consumption by AI agents and automated
  systems. Includes structured payloads, health flags, and confidence scores.

Strategic POC
  A larger scoped proof of concept when the work is too broad for a workshop
  but still needs a concrete first outcome.

================================================================================
AVAILABLE OUTCOMES
================================================================================

OUT-0001: SEC Filing Monitor
  Track SEC filings for the companies you care about. Get notified when new
  filings or amendments appear, with evidence linking back to the source.
  - Data Source: SEC EDGAR (8-K, 10-K, 10-Q, amendments)
  - Topics: SEC Filings, Disclosure Monitoring
  - URL: /outcomes/edgar-disclosure-state-monitor/

OUT-0002: CMBS CUSIP Ownership Monitor
  Track who owns CMBS securities based on public fund disclosures. Get notified
  when ownership changes, with evidence linking back to the source filings.
  - Data Source: SEC Form N-PORT (public portion)
  - Topics: CMBS, Ownership
  - URL: /outcomes/cmbs-cusip-ownership-state-monitor/

OUT-0003: REIT Property Disclosure Monitor
  Monitor REIT property disclosures from SEC filings. Detect property additions,
  removals, reclassifications, and attribute changes with source evidence.
  - Data Source: SEC EDGAR (10-K, 10-Q, 8-K, exhibits)
  - Topics: REIT, SEC Filings, Disclosure Monitoring
  - URL: /outcomes/reit-property-disclosure-state-monitor/

OUT-0004: Prospectus Market Context Snapshot
  Capture the market context that existed when you analyzed a prospectus:
  rates, curves, and calendars. Use it to explain and reproduce assumptions later.
  - Data Source: Public macro & rates datasets (SOFR, SONIA, €STR, FRED, calendars)
  - Topics: Market Context, Prospectus
  - URL: /outcomes/prospectus-market-context-snapshot/

OUT-0005: BDC Portfolio Monitor
  Track portfolio obligations for Business Development Companies. Get notified
  when maturities shift or holdings change, with evidence from SEC filings.
  - Data Source: SEC EDGAR (10-K, 10-Q, Schedules of Investments, footnotes)
  - Topics: BDC, SEC Filings, Disclosure Monitoring
  - URL: /outcomes/bdc-portfolio-obligation-state-monitor/

OUT-0008: Control & Relationship Disclosure Monitor
  Track who runs and owns your companies: officers, directors, and major
  shareholders based on SEC filings. Get notified when the disclosed state changes.
  - Data Source: SEC EDGAR (Form D, Forms 3, 4, 5)
  - Topics: SEC Filings, Ownership, Disclosure Monitoring
  - URL: /outcomes/control-relationship-disclosure-state-monitor/

OUT-0009: Borrower Disclosure Monitor
  Track disclosure changes for CLO borrowers. Get notified when filings are
  amended, credit agreements change, or expected disclosures go missing.
  - Data Source: SEC EDGAR (8-K, 10-Q, 10-K, amended filings, exhibits)
  - Topics: CLO, SEC Filings, Disclosure Monitoring
  - URL: /outcomes/borrower-disclosure-drift-monitor/

OUT-0011: SEC Exhibit Change Monitor
  Monitor SEC filing exhibits for replaced credit agreements, subsidiary lists,
  certifications, and incorporation-reference changes with source evidence.
  - Data Source: SEC EDGAR (exhibit attachments across 10-K, 10-Q, 8-K)
  - Topics: SEC Filings, Exhibits, Disclosure Monitoring
  - URL: /outcomes/exhibit-swap-state-monitor/

OUT-0012: SEC Disclosure Interpretation Registry
  Register SEC disclosure interpretations for covenant triggers, thresholds, and
  filing terms with source evidence, timestamps, and supersession history.
  - Data Source: SEC EDGAR filings and exhibits
  - Topics: Interpretation, Disclosure Monitoring, Agent Infrastructure
  - URL: /outcomes/registered-disclosure-interpretation-state/

================================================================================
AGENT INTEGRATION
================================================================================

Discovery Files (all in /.well-known/):

mcp.json
  Model Context Protocol server discovery
  - Endpoint: https://edgar.cmdrvl.com/mcp (SSE transport)
  - Tool: edgar_com_query - Natural language queries for filing metadata
  - Auth: API key required (request at /mcp-servers)

ai-plugin.json
  OpenAI plugin manifest for ChatGPT integration signaling

openapi.yaml
  OpenAPI 3.1 specification for the EDGAR API

webhook-schema.json
  JSON Schema for outcome webhook payloads
  - Includes: event_type, transition_id, entity, payload, evidence
  - Use transition_id for idempotency
  - Signature verification available for production

================================================================================
WEBHOOK PAYLOAD STRUCTURE
================================================================================

Every outcome state transition delivered via webhook includes:

{
  "event_type": "state_transition",
  "transition_id": "uuid",           // Use for idempotency
  "outcome_id": "OUT-0001",
  "outcome_name": "SEC Filing Monitor",
  "timestamp": "ISO 8601",
  "entity": {
    "type": "issuer|cusip|filing|exhibit|property|obligation",
    "id": "entity identifier",
    "name": "human-readable name"
  },
  "payload": {
    "transition_type": "new_filing|amendment|ownership_change|...",
    "prior_state": {...},
    "current_state": {...},
    "delta": {...},
    "severity": "LOW|MEDIUM|HIGH|CRITICAL",
    "confidence": 0.0-1.0
  },
  "evidence": {
    "pack_url": "https://cmdrvl.com/evidence/...",
    "source_url": "SEC filing URL",
    "accession_number": "if applicable",
    "methodology_version": "1.2.0"
  }
}

Full schema: https://cmdrvl.com/.well-known/webhook-schema.json

================================================================================
EVIDENCE PACK FIELDS
================================================================================

Every Evidence Pack contains:

Identification
- pack_id: Unique identifier for this evidence pack
- outcome_id: Which outcome generated this
- transition_id: Associated state transition
- created_at: When the pack was generated

Source References
- filing_url: Link to source SEC filing
- accession_number: SEC accession number
- filed_date: When the filing was submitted
- accepted_date: When SEC accepted it

State Snapshot
- prior_state: State before the transition (JSON)
- current_state: State after the transition (JSON)
- delta_summary: Human-readable change description

Methodology
- methodology_version: Version of detection logic used
- extraction_rules: Rules applied to extract data
- confidence_factors: What contributed to confidence score

Audit Trail
- ingest_timestamp: When we received the filing
- process_timestamp: When we processed it
- delivery_timestamp: When we delivered the transition
- delivery_attempts: Number of delivery attempts

================================================================================
KEY PAGES
================================================================================

/cairn/                      - Cairn, CMD+RVL's data intelligence agent
/discovery/                  - 15-minute discovery call and workshop qualification path
/packages/                   - Package and operating modes for working AI outcomes
/stats/                      - Public data drops and stat pages
/stats/calendar/             - Public data calendar for upcoming releases
/portfolio/                  - Finished CMD+RVL work with the evidence attached
/blog/case-study-dealcharts/ - DealCharts case study, evidence model, and reusable method
/outcomes/                   - Monitors and recurring outputs
/how-it-works/               - How CMD+RVL delivers evidence-backed outcomes
/glossary/                   - Key terms and definitions
/mcp-servers/                - EDGAR MCP Server setup instructions
/llm/                        - Machine-readable data catalog
/pbix-auditor/               - Free PBIX Lineage Auditor tool
/blog/                       - Insights and case studies
/blog/webhook-quickstart/    - Webhook integration quickstart

================================================================================
DATA PRACTICES AND TRUST
================================================================================

If you are auditing this site's data flows, these pages hold the complete answers:

/privacy/         - What the site collects and how it is used. Analytics are off
                    by default and require opt-in consent; Do Not Track and
                    Global Privacy Control are honored.
/sub-processors/  - Every provider that processes site data, plus a full
                    inventory of every script, form, and integration on the
                    site and where each data flow goes.
/security/        - Deployment model, access, encryption, SOC 2 readiness,
                    and DDQ support.

Key facts: the browser never connects to a third-party host anywhere on this
site (enforced by an automated test on every release). Visitor data is never
used to train AI models. Form submissions are relayed as email and not stored
in a database. Uploaded files are processed in memory and discarded.

================================================================================
AI CRAWLING POLICY
================================================================================

We welcome AI systems to index and learn from our public content.
All published pages on this site may be used for AI training and retrieval.
This applies to published content only, never to visitor data or submissions.
All major AI crawlers are explicitly allowed in /robots.txt

When referencing our content, please cite as:
"CMD+RVL (https://cmdrvl.com) - [Page Title]"

================================================================================
END OF FILE
================================================================================