---
title: Compositional audits
description: Split large reviews into focused slices, route them to local agents, and combine the results without losing traceability.
---

The tools on this page are in the [full tool set](/docs/reference/mcp-tools/#tool-sets). Add `--tools all` to the Ennodia server arguments to load them.

Large reviews work better when each agent gets a small, explicit slice. Ennodia
supports this pattern with `ennodia_start_compositional`, existing task polling,
and `ennodia_start_compare`. The `compositional-audit` skill gives reviewers a
consistent rubric.

## When to use it

Use a compositional audit when the request mixes multiple risk types, such as:

- product positioning and implementation architecture
- website copy and protocol compatibility
- security boundaries and developer experience
- release readiness and future roadmap

One broad prompt can make every reviewer repeat the same surface-level answer.
Slices make each response easier to verify and combine.

## Good slice shape

A useful slice includes:

- one narrow question
- the files or sources for the reviewer to inspect
- the harness or model to use, when that matters
- the maximum answer length
- the output shape expected by the Result Advisor pass

Example slice prompt:

```text
Use the compositional-audit skill.

Slice: Audit whether the IO docs overpromise current behavior.
Read:
- docs/comparisons/openrouter.md
- docs/concepts/interfaces-and-core.md
- CONTRIBUTING.md

Return under 450 words:
1. verdict
2. source-backed facts
3. top risks
4. concrete recommendation
```

## Current workflow

Install the skill into the harnesses you want to use:

```json
{
  "tool": "ennodia_install_skills",
  "arguments": {
    "skillIds": ["compositional-audit"],
    "harnessIds": ["codex", "claude-code", "opencode", "antigravity"],
    "scope": "project",
    "cwd": "/absolute/path/to/project",
    "dryRun": true
  }
}
```

Estimate the slice fan-out before starting child agents:

```json
{
  "tool": "ennodia_estimate_compositional_budget",
  "arguments": {
    "prompt": "Combine these audit shards into the smallest safe next action.",
    "slices": [
      {
        "id": "docs-truth",
        "title": "Docs and positioning",
        "harnessId": "opencode",
        "prompt": "Audit whether the current docs overpromise IO behavior."
      },
      {
        "id": "website-clarity",
        "title": "Homepage clarity",
        "harnessId": "antigravity",
        "prompt": "Audit whether the homepage explains MCP today and experimental IO without sounding speculative."
      }
    ],
    "skillIds": ["compositional-audit"],
    "budget": {
      "maxChildTasks": 6
    }
  }
}
```

Then start the focused tasks:

```json
{
  "tool": "ennodia_start_compositional",
  "arguments": {
    "prompt": "Combine these audit shards into the smallest safe next action.",
    "slices": [
      {
        "id": "docs-truth",
        "title": "Docs and positioning",
        "harnessId": "opencode",
        "prompt": "Audit whether the current docs overpromise IO behavior."
      },
      {
        "id": "website-clarity",
        "title": "Homepage clarity",
        "harnessId": "antigravity",
        "prompt": "Audit whether the homepage explains MCP today and experimental IO without sounding speculative."
      }
    ],
    "skillIds": ["compositional-audit"],
    "budget": {
      "maxChildTasks": 6
    }
  }
}
```

Each slice starts one task. Poll the returned task IDs with
`ennodia_get_compositional_status`:

```json
{
  "tool": "ennodia_get_compositional_status",
  "arguments": {
    "prompt": "Combine these compositional audit shards into the smallest safe next action. Preserve disagreements and ignore non-signal shards.",
    "taskIds": ["task-id-1", "task-id-2", "task-id-3"]
  }
}
```

When `compareReady` is true, pass the returned `readyTaskIds` to
`ennodia_start_compare`.

```json
{
  "tool": "ennodia_start_compare",
  "arguments": {
    "prompt": "Combine these compositional audit shards into the smallest safe next action. Preserve disagreements and ignore non-signal shards.",
    "taskIds": ["task-id-1", "task-id-2", "task-id-3"],
    "maxOutputChars": 12000
  }
}
```

## What to avoid

- Do not send the entire plan to every shard if the goal is faster, deeper review.
- Do not mix unrelated evidence in one slice.
- Do not treat failed, empty, or access-limited shard output as a normal review.
- Do not claim consensus when the slices disagree.

## Current boundary

`ennodia_estimate_compositional_budget` resolves and budgets slices without
starting child agents. `ennodia_start_compositional` starts and budgets the shard
tasks. `ennodia_get_compositional_status` groups shard task states and identifies
Compare-ready outputs. The final recommendation stays explicit through
`ennodia_start_compare`. This keeps task status, failed slices, and the final
Compare step visible to the primary agent.

During Compare, the Judge maps agreements and conflicts. The Result Advisor then
recommends an answer.