Metadata-Version: 2.4
Name: csoai-defoneos-isr-mcp
Version: 1.0.0
Summary: CSOAI DEFONEOS ISR (Intelligence, Surveillance, Reconnaissance) — sovereign UK defence ISR pipeline: object detection + geolocation + BFT-signed ISR + Cesium GeoJSON feed
Author-email: "MEOK AI Labs (CSOAI LTD)" <team@meok.ai>
License: MIT
Project-URL: Homepage, https://csoai.org/defoneos/isr
Project-URL: Repository, https://github.com/CSOAI-ORG/csoai-defoneos-isr-mcp
Keywords: mcp,defoneos,csoai,isr,intelligence,surveillance,reconnaissance,geolocation,yolov8,openathena,cesium,bft-council,uk-defence
Classifier: Development Status :: 4 - Beta
Classifier: License :: OSI Approved :: MIT License
Classifier: Programming Language :: Python :: 3.10
Classifier: Programming Language :: Python :: 3.11
Classifier: Programming Language :: Python :: 3.12
Classifier: Topic :: Security
Classifier: Topic :: Scientific/Engineering :: Information Analysis
Requires-Python: >=3.10
Description-Content-Type: text/markdown
Requires-Dist: mcp>=1.0.0
Requires-Dist: pydantic>=2.0

# csoai-defoneos-isr-mcp

**CSOAI DEFONEOS — sovereign UK defence ISR (Intelligence, Surveillance, Reconnaissance) pipeline.**

> PART OF THE DEFONEOS SENSOR LAYER (`MEOK_DEFONEOS_ALIGNMENT_2026-06-27.md` v2.0).

Wraps YOLOv8-style object detection + OpenAthena-style line-of-sight geolocation + Kalman-lite
track fusion + BFT-signed ISR reports into a single sovereign UK-defence-grade MCP surface.

Each tool returns a stable audit `sigil` (16-char sha256 prefix) so every ISR result is
reproducible, append-only traceable, and re-derivable by external auditors.

A `BannedTermGate` (per the Mavis Phantom-Context Strip rule) refuses any prompt containing
severed brands (James Castle, CSGA, Terranova, defonos.io, Toronto Summit phantoms, etc.).
No override path.

---

## Tools (10)

| # | Tool | Purpose |
|---|------|---------|
| 1 | `isr_classify_target`        | Classify objects in ISR imagery (10 defence-relevant categories) |
| 2 | `isr_geolocate_target`      | Geolocate target from sensor lat/lon/alt + bearing + elevation |
| 3 | `isr_fuse_track`            | Fuse N observations into a single Kalman-lite track |
| 4 | `isr_generate_report`       | Generate a BFT-signed ISR mission report |
| 5 | `isr_cesium_feed`           | Produce Cesium-compatible GeoJSON feed |
| 6 | `isr_target_categories`     | List all supported defence-relevant target categories |
| 7 | `isr_priority_score`        | Score a target by threat × confidence × proximity + fratricide-risk |
| 8 | `isr_route_to_cop`          | Compute sensor → Common Operating Picture route |
| 9 | `isr_corroborate_sources`   | Corroborate a track across N independent sources |
| 10 | `isr_chain_of_custody`     | Append a custody entry to the ISR evidence chain (JSONL) |

---

## Install

```bash
cd csoai-defoneos-isr-mcp
pip install -e .
```

## Run

```bash
# As MCP server (stdio)
csoai-defoneos-isr-mcp
```

## Tests

```bash
PYTHONPATH=. python3 -m pytest -v
```

12 tests, all green in <1s.

## Defence Target Categories

`vehicle_light`, `vehicle_heavy`, `vehicle_armoured`, `drone_fixed_wing`, `drone_rotary`,
`personnel`, `structure`, `emitter`, `vehicle_motorcycle`, `vessel_small`.

## Audit Chain

Append-only JSONL at `~/.csoai_defoneos_isr_custody.jsonl`. Each entry hashes
itself plus the prior chain position.

## Alignment

Per `MEOK_DEFONEOS_ALIGNMENT_2026-06-27.md` v2.0 §IS — the ISR pillar of the
sensor layer. Sister to `csoai-defoneos-mcp` (CERTIFIES) and
`csoai-defoneos-swarm-mcp` (swarm control).
