# Orchestration Policy Enforcer

> Enforces workflow order, required artifacts, quality gates, and reproducibility rules for the orchestrator. Use when a workflow must be run in a disciplined, auditable, reproducible way.

- Skill: `alainlebret/orchestration-policy-enforcer` (Agent Skill)
- Install (CLI): `npx skillmds@latest add alainlebret/orchestration-policy-enforcer`
- Raw SKILL.md: https://api.skillmd.com/api/skills/alainlebret/orchestration-policy-enforcer/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Productivity
- Author: alainlebret (https://skillmd.com/u/alainlebret)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/alainlebret/orchestration-policy-enforcer

---


# Orchestration Policy Enforcer

## Goal

Act as the final gatekeeper before a workflow's output is considered validated. Verify that all rules from `shared/orchestration-policy/` and `shared/reproducibility-policy/` have been applied, and emit a structured compliance report.

## When to use

Run as the **last step** of any non-trivial workflow, after all skills have executed and all artefacts have been produced.

## Inputs

- `workflow-spec.json` — the authoritative list of required skills, artefacts, and quality gates
- `run-manifest.json` — what was actually produced
- `quality-report.json` — gate verification results
- `language-report.json` — language compliance results
- `execution-plan.json` — planned vs. actual skill execution

## Output contract

| File | Description |
|---|---|
| `orchestration-compliance-report.md` | Structured report: one section per policy rule, with verdict and evidence |

### Report structure

```markdown
# Orchestration Compliance Report — <workflow_id> — <run_id>

## Policy: Required artefacts
- mission.json: present ✓
- execution-plan.json: present ✓
- run-manifest.json: present ✓
- quality-report.json: present ✓
- [other expected artefacts]: present ✓ / MISSING ✗

## Policy: Quality gates
- QG1 mission-present: pass ✓ / fail ✗
- ...

## Policy: Language compliance
- overall_status: pass ✓ / review ✗

## Policy: Reproducibility
- run_id present: yes ✓
- created_at present: yes ✓
- skills_used matches execution_order: yes ✓ / MISMATCH ✗

## Overall verdict
COMPLIANT / NON-COMPLIANT — [summary of blocking issues]
```

## Compliance rules (from shared policies)

| Rule | Source | Check |
|---|---|---|
| `mission.json` must be present | orchestration-policy.md | `run-manifest.json → artifacts_out` contains `mission.json` |
| `execution-plan.json` must be present | orchestration-policy.md | file exists in run or example directory |
| `run-manifest.json` must be present | orchestration-policy.md | file exists with valid `run_id` |
| At least one audit artefact must exist | orchestration-policy.md | `quality-report.json` or `orchestration-compliance-report.md` present |
| Launcher must not silently invent inputs | launcher-policy.md | `artifacts_in` in manifest only lists files that actually existed |
| Quality gates must be stored as JSON | launcher-policy.md | `quality-report.json` exists and is valid JSON |
| `run_id` must be unique and timestamped | reproducibility-policy | `run_id` matches pattern `RUN-<WORKFLOW>-<ISO8601>` |
| Language report must be emitted | workflow language_constraints | `language-report.json` exists if `must_emit_language_report: true` |

## Rules

- Report all non-compliances, not just the first one found.
- A workflow with any blocking non-compliance must have `overall_status: fail`.
- A workflow with only warnings must have `overall_status: review`.
- A fully compliant workflow must have `overall_status: pass`.