Skill Overview

26 workflows use the agentic-workflows: MCP tool to analyze workflow runs, download logs, and inspect workflow status — but there is no shared documentation component for this tool. This is a significant gap compared to similar MCP tools: shared/mcp/serena-go.md exists and is used by 16 workflows to document Serena Go analysis patterns. Creating the analogous shared/mcp/agentic-workflows.md would standardize usage patterns across all 26 workflows.

Current Usage

The agentic-workflows: tool is used in:

• agent-performance-analyzer.md
• audit-workflows.md
• cloclo.md
• craft.md
• daily-cli-tools-tester.md
• daily-firewall-report.md
• daily-observability-report.md
• daily-rendering-scripts-verifier.md
• daily-safe-output-optimizer.md
• deep-report.md
• dev-hawk.md
• example-workflow-analyzer.md
• mcp-inspector.md
• metrics-collector.md
• portfolio-analyst.md
• prompt-clustering-analysis.md
• python-data-charts.md
• q.md
• safe-output-health.md
• security-review.md
• smoke-claude.md
• smoke-copilot-arm.md
• smoke-copilot.md
• static-analysis-report.md
• workflow-health-manager.md
• workflow-normalizer.md

Proposed Shared Component

File: .github/workflows/shared/mcp/agentic-workflows.md

Configuration:

---
# Agentic Workflows MCP Server
# Provides tools for analyzing and managing agentic workflow runs
#
# Usage:
#   imports:
#     - shared/mcp/agentic-workflows.md

tools:
  agentic-workflows:
---

## Agentic Workflows MCP Server

The `agentic-workflows` MCP server provides tools for analyzing workflow execution history, downloading logs, and auditing runs.

### Available Tools

#### `status` — List Workflow Runs
Lists recent workflow runs with their status, engine, and execution metadata.

#### `audit` — Deep Run Analysis  
Retrieves detailed execution data for a specific run including tool calls, agent output, and token usage.

#### `logs` — Download Execution Logs
Downloads complete execution logs for workflow runs. Use with the `--start-date`, `--engine`, and `--json` flags for targeted analysis.

### Common Analysis Patterns

```bash
# List recent workflow runs
# Tool: status with { "workflow": "my-workflow", "limit": 10 }

# Analyze a specific run
# Tool: audit with { "run_id": "12345" }

# Batch log analysis
# Use the steps: section with ./gh-aw logs for pre-fetching

**Usage Example**:
```yaml
# In a workflow
imports:
  - shared/mcp/agentic-workflows.md

tools:
  agentic-workflows:  # Included via import; no need to repeat

Impact

• Workflows affected: 26 workflows
• Lines saved: Primarily documentation + consistency; ~10–15 lines of boilerplate usage instructions per workflow = ~260–390 lines of duplicated inline docs
• Maintenance benefit: Single source of truth for agentic-workflows tool usage; easier to keep in sync with tool API changes
• Onboarding: New workflow authors can immediately understand how to use the tool correctly by importing this component

Implementation Plan

1. Study existing shared/mcp/serena-go.md for structural reference
2. Create shared/mcp/agentic-workflows.md with full usage documentation
3. Document all available MCP tools (status, audit, logs) with examples
4. Identify which of the 26 workflows contain inline usage instructions that could be removed
5. Refactor 3–5 high-priority workflows to import the component (start with audit-workflows.md and workflow-health-manager.md)
6. Verify compiled lock files are valid after changes

Related Analysis

This recommendation comes from the Workflow Skill Extractor analysis run on 2026-03-08.

Generated by Workflow Skill Extractor · ◷

• expires on Mar 10, 2026, 4:24 PM UTC