# Coordinate Evidence Review Agents

Complex evidence workflows can lose context, citations, and ownership across multiple agents. This skill creates a plan-first coordination layer for tests, delegation, audits, and quality control.

## Install

```bash
npx skillstore add acskamloops/manager-planner
```

## Metadata

- - Status: approved
- - Slug: acskamloops-manager-planner
- - Version: 1.0.0
- - Author: ACSKamloops
- - GitHub username: ACSKamloops
- - License: MIT
- - Repository: https://github.com/ACSKamloops/shs-engine/tree/master/.codex/skills/manager-planner
- - Ref: master
- - Supported tools: Claude, Codex, Claude Code
- - Risk level: medium
- - Risk factors: external\_commands
- - Quality score: 73
- - Public page: https://skillstore.pages.dev/skills/acskamloops-manager-planner
- - Manifest: https://skillstore.pages.dev/api/skills/acskamloops-manager-planner/manifest

## Capabilities

- Requires a written plan before major workflow actions.
- Routes work to named Pukaist agent roles by task type.
- Directs agents to run repository integrity and health checks.
- Defines checks for logs, queues, citations, and output quality.
- Applies a neutral clerk standard for evidence reporting.

## Use Cases

- Plan a New Evidence Intake: Create an approved sequence for ingestion, OCR, review, and logging before any agent changes files.
- Audit a Review Pipeline: Check queues, logs, citations, and refined reports for consistency before marking work complete.
- Coordinate Specialist Agents: Assign work to Gatekeeper, Scribe, Analyst, Archivist, Historian, or Barrister roles based on the task.

## Prompt Templates

### Create a Work Plan

```
Use manager-planner to draft a step-by-step plan for this evidence workflow. Do not execute the plan until I approve it.
```

### Check System Health

```
Use manager-planner to check whether the evidence workflow is healthy. Summarize required tests, log checks, and queue checks before running commands.
```

### Delegate Agent Work

```
Use manager-planner to assign this request to the correct Pukaist agent roles. Include acceptance criteria for each delegated task.
```

### Audit Final Outputs

```
Use manager-planner to perform a final quality audit across refined reports, citations, queue state, and communication logs. Report blockers first.
```

## Limitations

- Depends on repository files and scripts that may not exist elsewhere.
- Can only coordinate agents when the host tool supports delegation.
- Does not perform OCR, analysis, or legal synthesis by itself.
- Requires human approval before delegated execution.

## Best Practices

- Ask for a plan before allowing agents to modify files.
- Run integrity checks before major operations or approvals.
- Verify citations and queue status before marking work complete.

## Anti Patterns

- Using the skill to bypass user approval for implementation.
- Running cleanup or reconciliation commands without reviewing scope.
- Treating agent summaries as evidence without checking citations.

## Security Audit

- - Safe to publish: true
- - Audited at: 2026-06-28T03:51:25.132\+00:00
- - Summary: The static analyzer flagged many Markdown backticks as Ruby or shell execution, and those are mostly false positives. Human review confirmed that the skill does instruct agents to run repository-local Python utilities and cleanup commands, so it carries a legitimate external-command risk but no evidence of malicious intent.

## Stats

- - Views: 196
- - Downloads: 6
- - Favorites: 0
- - Popularity score: 0
