IG Analyze -- API-driven Instagram performance analysis

Key references:

• references/account-baseline.md -- current benchmarks, follower count, historical averages
• references/scoring-system.md -- metric definitions, scoring thresholds, weighted formulas

CRITICAL RULE: This skill MUST use live API data. Never guess, estimate, or fabricate metrics. If an API call fails, report the failure. Do not fill in placeholder numbers.

Phase 1: Data Collection

Pull data using Instagram MCP tools in this order:

1. Account overview:

   mcp__claude_ai_Instagram_MCP__get_user_info

   Capture: follower count, following count, media count, bio, profile pic.
2. Recent media:

   mcp__claude_ai_Instagram_MCP__get_user_media

   Fetch the most recent 25 posts. For each, capture: id, type (carousel/reel/image), timestamp, caption, like count, comment count.
3. Post-level insights (for each post):

   mcp__claude_ai_Instagram_MCP__get_post_insights

   Capture: reach, impressions, saves, shares, video views (if reel), profile visits.
4. Account-level insights:

   mcp__claude_ai_Instagram_MCP__get_user_insights

   Capture: reach, impressions, follower growth, profile visits over the requested period.

Bundle independent API calls in parallel where possible. Do not make serial calls when parallel is available.

Phase 2: Metric Calculation

Calculate the following metrics from the collected data:

Per-Post Metrics

Account-Level Metrics

Round all percentages to 2 decimal places.

Phase 3: Pattern Detection

Analyze the calculated metrics to identify patterns:

1. Best performers: Top 3 posts by engagement rate. Note format, hook category, posting day/time.
2. Worst performers: Bottom 3 posts. Note format, hook category, posting day/time.
3. Format comparison: Average engagement rate per format (reel, carousel, image).
4. Hook category analysis: If hook-library.md categories are identifiable in captions, group performance by category.
5. Posting time analysis: Group by day of week and hour. Identify best and worst windows.
6. Save rate leaders: Top 3 posts by save rate (these indicate high-value content).

Phase 4: Benchmark Comparison

Load references/account-baseline.md and compare current metrics:

Flag any metric that is more than 20% below baseline as a warning.

Phase 5: Quality Check

Before delivering the report, verify:

• All metrics are based on actual API data (no estimates)
• Per-post and account-level metrics are both included
• At least one actionable insight per pattern detected
• Benchmark comparison uses current account-baseline.md values
• Numbers are internally consistent (engagement rate matches component metrics)

Phase 6: Report Output

Deliver as a structured Markdown report:

## Instagram Performance Report

**Period:** [date range]
**Posts analyzed:** [count]
**Follower count:** [current]

### Key Metrics (Account Average)
| Metric | Value | vs Baseline |
|--------|-------|-------------|
| Engagement Rate | X% | +/-Y% |
| Save Rate | X% | +/-Y% |
| Reach-to-Follower | X% | +/-Y% |

### Top Performers
1. [Post description] -- [engagement rate] -- [why it worked]
2. ...
3. ...

### Format Breakdown
| Format | Avg Engagement | Avg Save Rate | Count |
|--------|---------------|---------------|-------|

### Patterns Detected
- [pattern 1 with actionable recommendation]
- [pattern 2 with actionable recommendation]

### Warnings
- [any metric >20% below baseline]

### Recommendations
1. [specific, actionable recommendation]
2. [specific, actionable recommendation]
3. [specific, actionable recommendation]

If the user requests a specific post analysis (single post), skip account-level metrics and deliver only per-post analysis with benchmark comparison.