Contested AI Model Governance and Mission Assurance Cell
Mission Scope
- Treat this skill as a planning and decision-support aid for U.S. warfighter missions in its domain.
- Confirm echelon, operating environment, authorities, time horizon, and decision points before analysis.
- Keep products unclassified by default unless the user provides handling guidance and controlled data.
Workflow
- Frame the mission problem with command intent, constraints, and assumptions.
- Build one recommended option and at least two alternatives with explicit tradeoffs.
- Bind recommendations to joint, interagency, and coalition dependencies where applicable.
- Tie recommendations to observable triggers, branch conditions, and revalidation checks.
- Publish commander-readable and staff-action products with owners and suspense.
Required Output Format
Deliver in this order:
- Situation snapshot.
- Recommended option.
- Alternative options.
- Decision points and triggers.
- Staff tasking with owners and suspense.
Domain Products
Primary products for this skill: model trust posture report, rollback trigger matrix, mission assurance control plan.
External Tools and Protocol Integration
- Use
../_shared/references/external-tools-protocols.md and select a mission-fit profile from ../_shared/references/joint-operations-external-toolchain-profiles.md.
- Select concrete tool suites from
../_shared/references/warfighter-external-tool-and-protocol-catalog.md.
- Include one primary system-of-record and one cross-check source for critical outputs.
- State protocol/transport for outbound coordination (
USMTF, VMF, Link 16 J-series, CoT, STIX/TAXII, OGC, AIS/NMEA, HL7/FHIR, or coalition formats).
- Include provenance metadata in outputs: source system, refresh time (UTC), assumptions, and confidence.
Tool Invocation Contract
For each critical tool recommendation, include objective, required inputs, query/action template, expected schema, protocol/transport, and fallback path.
Mission Tool Authority Gates
- Apply escalation requirements in
../_shared/references/warfighter-tool-authority-gates.md for high-consequence recommendations.
- Include
authority_tier, decision_impact_level, approval_role, and audit_record_id for recommendations that can alter mission posture.
- If authority, legality, or provenance is uncertain, downgrade to advisory-only and request human command decision.
Interoperability Validation Checklist
- Run
../_shared/references/mission-assurance-checklist.md before final release.
- Validate provenance, protocol format, UTC freshness, confidence, and known gaps.
- If checks fail, provide a degraded-mode plan with explicit risk effects.
Guardrails
- Flag assumptions that exceed evidence.
- Separate facts, assessed judgments, and unknowns.
- Identify legal, policy, ROE, safety, and coalition constraints early.
- Do not provide weapon-employment procedures or bypasses to safeguards.
1---2name: contested-ai-model-governance-and-mission-assurance-cell3description: Govern AI model deployment, rollback, and trust posture in contested operations. Use when mission AI outputs influence command decisions and require rapid governance under adversarial pressure.4---56# Contested AI Model Governance and Mission Assurance Cell78## Mission Scope910- Treat this skill as a planning and decision-support aid for U.S. warfighter missions in its domain.11- Confirm echelon, operating environment, authorities, time horizon, and decision points before analysis.12- Keep products unclassified by default unless the user provides handling guidance and controlled data.1314## Workflow15161. Frame the mission problem with command intent, constraints, and assumptions.172. Build one recommended option and at least two alternatives with explicit tradeoffs.183. Bind recommendations to joint, interagency, and coalition dependencies where applicable.194. Tie recommendations to observable triggers, branch conditions, and revalidation checks.205. Publish commander-readable and staff-action products with owners and suspense.2122## Required Output Format2324Deliver in this order:25261. Situation snapshot.272. Recommended option.283. Alternative options.294. Decision points and triggers.305. Staff tasking with owners and suspense.3132## Domain Products3334Primary products for this skill: model trust posture report, rollback trigger matrix, mission assurance control plan.3536## External Tools and Protocol Integration3738- Use `../_shared/references/external-tools-protocols.md` and select a mission-fit profile from `../_shared/references/joint-operations-external-toolchain-profiles.md`.39- Select concrete tool suites from `../_shared/references/warfighter-external-tool-and-protocol-catalog.md`.40- Include one primary system-of-record and one cross-check source for critical outputs.41- State protocol/transport for outbound coordination (`USMTF`, `VMF`, `Link 16 J-series`, `CoT`, `STIX/TAXII`, `OGC`, `AIS/NMEA`, `HL7/FHIR`, or coalition formats).42- Include provenance metadata in outputs: source system, refresh time (UTC), assumptions, and confidence.4344## Tool Invocation Contract4546For each critical tool recommendation, include objective, required inputs, query/action template, expected schema, protocol/transport, and fallback path.4748## Mission Tool Authority Gates4950- Apply escalation requirements in `../_shared/references/warfighter-tool-authority-gates.md` for high-consequence recommendations.51- Include `authority_tier`, `decision_impact_level`, `approval_role`, and `audit_record_id` for recommendations that can alter mission posture.52- If authority, legality, or provenance is uncertain, downgrade to advisory-only and request human command decision.5354## Interoperability Validation Checklist5556- Run `../_shared/references/mission-assurance-checklist.md` before final release.57- Validate provenance, protocol format, UTC freshness, confidence, and known gaps.58- If checks fail, provide a degraded-mode plan with explicit risk effects.5960## Guardrails6162- Flag assumptions that exceed evidence.63- Separate facts, assessed judgments, and unknowns.64- Identify legal, policy, ROE, safety, and coalition constraints early.65- Do not provide weapon-employment procedures or bypasses to safeguards.
zwright8