Shipkit Metrics

shipkit-metrics - Metric Capture Utility

Purpose: Let users record current metric values into .shipkit/metrics/latest.json so that /shipkit-product-goals --evaluate and /shipkit-engineering-goals --evaluate can compare actuals against targets.

This is a manual capture tool — metrics come from real-world sources (analytics, APMs, test runs, user feedback) that no skill can auto-generate.

---

When to Invoke

User triggers:

• "Log metrics", "Update metrics", "Record measurements"
• "Our response time is now 200ms"
• "Conversion rate hit 12%"
• "Test coverage is at 85%"

Workflow position:

• After goals are defined (/shipkit-product-goals, /shipkit-engineering-goals)
• Before goal evaluation (--evaluate mode on goals skills)

---

Prerequisites

Recommended (to know which metrics to capture):

• .shipkit/goals/product.json — P-* criteria with thresholds
• .shipkit/goals/engineering.json — E-* criteria with thresholds

If missing: Accept any metric the user provides, using freeform IDs.

---

Arguments

• No args or metric text → Interactive mode (show goals, ask for values)
• --list → Show current metrics and which criteria are still unmeasured
• Metric ID + value (e.g., P-001 82%) → Quick capture mode

---

Process

Mode 1: Interactive (default)

1. Read .shipkit/goals/product.json and .shipkit/goals/engineering.json if they exist
2. Extract all criteria IDs and their metrics/thresholds
3. Read .shipkit/metrics/latest.json if exists (to show what's already captured)
4. Present unmeasured criteria:

Metrics Capture — {N} criteria defined, {M} measured

Unmeasured:
  P-001: Task completion rate (target: > 80%)
  P-003: Page load satisfaction (target:  80%) — exceeded
  P-002: Error rate = 2.1% (target:  80% — exceeded)`

### Mode 3: List (`--list`)

1. Read goals files and metrics file
2. Show all criteria with current values and status
3. Highlight unmeasured criteria

---

## Write Strategy

**READ-MODIFY-WRITE** on `.shipkit/metrics/latest.json`:
- Read existing file (or create if missing)
- Add/update metric entries
- Update `lastUpdated` timestamp
- Update `summary` counts
- Write complete file back

---

## Context Files This Skill Reads

| File | Purpose | If Missing |
|------|---------|------------|
| `.shipkit/goals/product.json` | P-* criteria and thresholds | Accept freeform metrics |
| `.shipkit/goals/engineering.json` | E-* criteria and thresholds | Accept freeform metrics |
| `.shipkit/metrics/latest.json` | Existing measurements | Create new |

## Context Files This Skill Writes

| File | When |
|------|------|
| `.shipkit/metrics/latest.json` | Created on first run, updated on subsequent runs |

---

## When This Skill Integrates with Others

### Before This Skill
| Skill | Why |
|-------|-----|
| `shipkit-product-goals` | Defines P-* criteria with thresholds |
| `shipkit-engineering-goals` | Defines E-* criteria with thresholds |

### After This Skill
| Skill | How |
|-------|-----|
| `shipkit-product-goals --evaluate` | Reads metrics/latest.json to compare actuals vs targets |
| `shipkit-engineering-goals --evaluate` | Same |
| `shipkit-stage --evaluate` | Cross-references all criteria for graduation evidence |

---

## After Completion

**Suggest next**: Run `/shipkit-product-goals --evaluate` or `/shipkit-engineering-goals --evaluate` to see how actuals compare to targets.

## Success Criteria

- [ ] At least one metric value recorded
- [ ] File saved to `.shipkit/metrics/latest.json`
- [ ] Each entry has criterionId, value, measuredAt, and source
- [ ] Summary counts match actual entries

## Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

- **Author:** [stefan-stepzero](https://github.com/stefan-stepzero)
- **Source:** [stefan-stepzero/shipkit](https://github.com/stefan-stepzero/shipkit)
- **License:** MIT

Install and usage instructions live in the source repository linked above.