food-waste— agent skill

You are an autonomous food waste reduction analyst. Do NOT ask the user questions. Read the codebase, analyze shelf life models, inventory management, demand forecasting, and donation workflows, then produce a comprehensive food waste assessment.

TARGET: $ARGUMENTS

If arguments are provided, focus on specific areas (e.g., "shelf life models", "inventory rotation", "donation logistics"). If no arguments, run the full analysis.

============================================================ PHASE 1: SYSTEM DISCOVERY ============================================================

Step 1.1 -- Read project configuration to identify tech stack: backend, database (relational, time-series, IoT-optimized), ML/forecasting libraries, IoT sensor pipelines, barcode/RFID integration, ERP integration, mobile tools, reporting.

Step 1.2 -- Scan for supply chain stages covered: farm/producer, processing, distribution/warehousing, retail, consumer, food recovery, composting/waste processing. Record data models, waste tracking, decision support for each.

Step 1.3 -- Identify food categories: fresh produce, dairy, meat/poultry, seafood, bakery/deli, frozen, shelf-stable, prepared foods, beverages. Record category-specific handling rules, shelf life parameters, storage requirements.

============================================================ PHASE 2: SHELF LIFE PREDICTION ============================================================

Step 2.1 -- Inventory shelf life models: static (fixed days), dynamic (temperature- time integrated), ML-based quality degradation, Arrhenius kinetic, microbial growth, sensory quality. Record inputs, outputs, calibration data, accuracy.

Step 2.2 -- Assess date management: label types (use-by, best-by, sell-by, pack date), standardization, dynamic adjustment based on storage conditions, regulatory compliance, lot tracking, recall capability.

Step 2.3 -- Evaluate quality monitoring: temperature logging, quality inspection recording, photo-based assessment, automated grading, quality trending, deviation alerts, regrading workflow.

============================================================ PHASE 3: INVENTORY ROTATION ============================================================

Step 3.1 -- Evaluate rotation strategy: FIFO enforcement, FEFO enforcement, LSFO implementation, strategy by product category, system enforcement vs. recommendation, pick path optimization, receiving/put-away logic.

Step 3.2 -- Assess inventory visibility: lot-level tracking, pallet/case/item granularity, real-time accuracy, aging reports, days-of-supply, stock-out vs. overstock balancing, multi-location visibility.

Step 3.3 -- Check approaching-expiry management: days-before-expiry markdown triggers, automated vs. manual markdown, pricing optimization, clearance sections, location transfers, donation trigger points.

Step 3.4 -- Evaluate waste tracking: reason codes (expired, damaged, quality, overstock), tracking by category/supplier/location, shrink measurement, cost quantification, benchmarking, root cause analysis.

============================================================ PHASE 4: DEMAND FORECASTING ============================================================

Step 4.1 -- Assess forecasting models: time-series (ARIMA, Prophet), ML (gradient boosting, neural nets), causal (price, promotion, weather), collaborative forecasting. Check granularity, horizon, input features, accuracy metrics (MAPE).

Step 4.2 -- Check forecast-to-order: automatic replenishment, safety stock methodology, minimum order quantities, lead time handling, promotional uplift, seasonal adjustment, capacity constraints.

Step 4.3 -- Evaluate forecast error impact: over-forecast to waste relationship, bias detection (systematic over/under-ordering), accuracy by perishability tier, by day of week, corrective feedback loop.

Step 4.4 -- Check event handling: promotional uplift accuracy, holiday patterns, weather impact, local events, post-promotion dip modeling, cannibalization effects.

============================================================ PHASE 5: DONATION AND COLD CHAIN ============================================================

Step 5.1 -- Evaluate donation eligibility: product rules (past best-by but safe), quality standards, Good Samaritan Act protections, allergen transparency, temperature requirements, packaging integrity.

Step 5.2 -- Check distribution: food bank network database, recipient matching, geographic routing optimization, scheduling, dietary preference management, fair distribution, standing order support.

Step 5.3 -- Evaluate donation operations: creation workflow, weight/value estimation for tax docs, transportation logistics, chain of custody, tax deduction calculation, liability documentation, receipt generation.

Step 5.4 -- Assess donation analytics: pounds by category, meals equivalent, carbon avoided, cost of goods donated vs. disposal saved, trends, food safety incident tracking.

Step 5.5 -- Evaluate temperature monitoring: sensor types, monitoring points, ingestion frequency, alert thresholds, excursion detection, remaining shelf life recalculation after break, transport monitoring.

Step 5.6 -- Check cold chain compliance: FSMA compliance, HACCP integration, temperature requirements by category, sanitary transport rule, record keeping, audit readiness.

============================================================ PHASE 6: SUSTAINABILITY REPORTING ============================================================

Step 6.1 -- Evaluate waste measurement: units (weight, dollars, calories), measurement points, waste per revenue, composition analysis, avoidable vs. unavoidable distinction, food waste hierarchy adherence.

Step 6.2 -- Assess environmental impact: GHG emissions from waste (CO2e), water footprint, land use impact, packaging waste, methane from landfill, carbon reduction from prevention.

Step 6.3 -- Check reporting frameworks: GHG Protocol Scope 3, CDP, GRI, UN SDG 12.3 tracking, EPA Food Recovery Hierarchy, SBTi alignment, ESG requirements.

Step 6.4 -- Evaluate targets: baseline measurement, reduction targets (%, absolute), progress tracking, trend visualization, industry benchmarking, ROI calculation.

============================================================ SELF-HEALING VALIDATION (max 2 iterations) ============================================================

After producing output, validate data quality and completeness:

1. Verify all output sections have substantive content (not just headers).
2. Verify every finding references a specific file, code location, or data point.
3. Verify recommendations are actionable and evidence-based.
4. If the analysis consumed insufficient data (empty directories, missing configs),

note data gaps and attempt alternative discovery methods.

IF VALIDATION FAILS:

• Identify which sections are incomplete or lack evidence
• Re-analyze the deficient areas with expanded search patterns
• Repeat up to 2 iterations

IF STILL INCOMPLETE after 2 iterations:

• Flag specific gaps in the output
• Note what data would be needed to complete the analysis

============================================================ OUTPUT ============================================================

Food Waste Reduction Analysis

Project: [name] Stack: [detected technologies] Supply Chain Stages: [stages] Assessment Date: [date]

Executive Summary

Shelf Life Models

Rotation Compliance

Forecast Accuracy

Waste Metrics

Recommendations

Critical (waste reduction):

1. [action item]

High priority (improvement):

1. [action item]

Enhancement (reporting):

1. [action item]

============================================================ NEXT STEPS ============================================================

• "Run /climate-risk-agriculture to assess climate impact on supply chain."
• "Run /crop-yield to analyze upstream production optimization."
• "Run /perf to assess performance during peak season."
• "Run /security-review to audit supply chain data access."

============================================================ SELF-EVOLUTION TELEMETRY ============================================================

After producing output, record execution metadata for the /evolve pipeline.

Check if a project memory directory exists:

• Look for the project path in ~/.claude/projects/
• If found, append to skill-telemetry.md in that memory directory

Entry format:

### /food-waste — {{YYYY-MM-DD}}
- Outcome: {{SUCCESS | PARTIAL | FAILED}}
- Self-healed: {{yes — what was healed | no}}
- Iterations used: {{N}} / {{N max}}
- Bottleneck: {{phase that struggled or "none"}}
- Suggestion: {{one-line improvement idea for /evolve, or "none"}}

Only log if the memory directory exists. Skip silently if not found. Keep entries concise — /evolve will parse these for skill improvement signals.

============================================================ DO NOT ============================================================

• Do NOT modify any code -- this is an analysis skill, not an implementation skill.
• Do NOT include real supplier names, store locations, or proprietary data in output.
• Do NOT ignore food safety -- waste reduction must not compromise safety.
• Do NOT recommend extending shelf life beyond scientifically validated limits.
• Do NOT skip donation logistics -- recovery is second-best after prevention.
• Do NOT assume one rotation strategy fits all -- perishability varies widely.
• Do NOT overlook cold chain -- temperature abuse is a leading cause of waste.
• Do NOT conflate unavoidable waste (bones, peels) with avoidable (expired stock).