evaluate

Process

Step 1: Determine Target

Parse the input to determine evaluation target:

• Path argument provided → local project (e.g., /path/to/project)
• Name or URL argument provided → external app (e.g., "Bear", "Notion")
• No argument → local project at current working directory

If the target is a local project, read its README (or equivalent top-level docs) to understand what it does. If no README exists or the product purpose is unclear, ask the user for a one-sentence product description.

Step 2: Detect Platform

Local project — check for platform indicators:

• .xcodeproj, .xcworkspace, Package.swift with iOS platform → iOS
• package.json, next.config, vite.config → Web
• pubspec.yaml → Flutter (cross-platform)
• android/ directory → Android
• Ambiguous → ask the user

External app:

• If user specified platform, use it
• If app name is well-known, infer (but confirm if uncertain)
• Otherwise, ask

Step 3: Resolve Reference Paths

Locate the plugin's reference files by searching for **/product-lens/references/_calibration.md. From the same parent directory, resolve absolute paths to:

• _calibration.md (always)
• _scoring.md (always)
• dimensions/01-demand-authenticity.md through dimensions/06-execution-quality.md (all 6)
• modules/kill-criteria.md, modules/feature-audit.md, modules/elevator-pitch.md, modules/pivot-directions.md, modules/validation-playbook.md (all 5)

Step 4: Gather Market Context

Dispatch the market-scanner agent with:

• Product description (from README or user input)
• Target category (inferred from product description)
• Known competitors (if user mentioned any)
• Platform (from Step 2)

Wait for market-scanner to complete before proceeding. The dimension evaluators need this data.

Step 5: Pre-merge Sub-Questions

Read _calibration.md once — this will be injected as preamble into every dimension-evaluator prompt.

Determine the platform variant to use:

• iOS detected → extract ### iOS sections from each dimension file
• Otherwise → extract ### Default sections

For each of the 6 dimension files, read the file and extract:

1. Core question (the line after **Core question:**)
2. Universal sub-questions (the numbered list under ## Universal Sub-Questions)
3. Platform-specific sub-questions (the numbered list under the correct ### [Platform] section within ## Platform-Specific Sub-Questions). If iOS, also extract the iOS core question variant (the blockquote under the ### iOS heading).
4. Scoring anchors (the table under ## Scoring Anchors)
5. Evidence sources (the content under ## Evidence Sources)

For each dimension, merge the universal + platform-specific sub-questions into a single numbered list (universal questions first, then platform-specific, numbered sequentially).

Result: 6 self-contained dimension payloads.

Also prepare market data excerpts per dimension from the market-scanner output:

• Demand Authenticity: full market data
• Journey Completeness: "none" (code analysis dimension)
• Market Space: Direct Competitors + Market Signals + Pricing Landscape sections
• Business Viability: Pricing Landscape + Discovery Channels sections
• Moat: Risk Signals + Direct Competitors sections
• Execution Quality: "none" (code analysis dimension)

Step 6: Dispatch 6x dimension-evaluator (parallel)

Dispatch all 6 dimension-evaluator agents in a single message (parallel execution). Each agent receives:

• Calibration context: full content of _calibration.md
• Dimension name (English + Chinese) and core question
• Sub-questions: the merged numbered list from Step 5
• Scoring anchors: the table from Step 5
• Evidence source hints: from Step 5
• Product info: name, one-sentence description, evaluation type (local/external), project root path
• Market data excerpt: the dimension-relevant excerpt from Step 5
• Depth mode: standard

Wait for all 6 to complete.

Step 7: Collect and Validate Dimension Results

For each of the 6 returned results, verify:

1. Section header ## [Dimension Name exists
2. Score line contains star characters (★/☆), not X/10 or other scales
3. Count of ### Q sub-sections matches the expected sub-question count for that dimension
4. Each sub-section contains **Evidence:** and **Assessment:** fields
5. **Anchor match:** field exists in the Dimension Score section
6. ### Next Action section exists after ### Dimension Score

If any dimension fails validation:

• Re-dispatch that single dimension-evaluator with a correction note prepended to the original prompt: "Your previous output had these issues: [list specific failures]. Produce the corrected output following the template exactly."
• Maximum 1 retry per dimension
• If still non-compliant after retry: include with warning annotation > ⚠️ This dimension's output did not fully comply with the evaluation template.

Extract from each valid result:

• Dimension star score (for the overview table and extras-generator)
• One-sentence justification (for the overview table)
• Next Action text (for Priority Actions section in the final report)

Step 8: Dispatch extras-generator

Read the 5 module files from references/modules/. For each module file that has a ## Platform Additions or ## Platform Constraints section:

• If iOS: extract the ### iOS section and append it to the module's base instructions
• Otherwise: use the base instructions only

Dispatch the extras-generator agent with:

• Product info: name, description, evaluation type, project root path, platform
• Dimension scores: all 6 dimension names with their star scores and one-sentence justifications
• Weak dimensions: list of dimensions scored <=2
• Dimensions scored <=3: list of dimensions scored <=3, with their Next Action text (for Validation Playbook)
• Module instructions: the 5 module files' content (pre-merged with platform additions)
• Market data excerpt: full market-scanner output (for Pivot Directions context)

Wait for completion.

Step 9: Assemble Final Report

Read _scoring.md. Compute the weighted total from the 6 dimension scores using default equal weights (or a user-specified preset if requested).

Assemble the final report by combining dimension results and extras output:

# Product Lens: [Product Name]

> [One-sentence product description]

## Elevator Pitch Test
[from extras-generator output — the Elevator Pitch Test section]

## Evaluation Overview

| Dimension | Score | Justification |
|-----------|-------|---------------|
| Demand Authenticity (需求真伪) | [stars] | [justification from dimension result] |
| Journey Completeness (逻辑闭环) | [stars] | [justification] |
| Market Space (市场空间) | [stars] | [justification] |
| Business Viability (商业可行) | [stars] | [justification] |
| Moat (护城河) | [stars] | [justification] |
| Execution Quality (执行质量) | [stars] | [justification] |
| **Weighted Total** | **X.X** | |

## Priority Actions

[Ordered list of Next Actions from dimensions scored ≤3★, sorted by score ASC (lower = more urgent).
Each item: "**[Dimension Name] (N★):** [Next Action text from dimension-evaluator output]"
If no dimensions scored ≤3★, write: "All dimensions scored ≥4★. See individual dimension Next Actions for consolidation opportunities."]

## Dimension Details

[All 6 dimension evaluation results in order, each preserving its internal structure
 (## heading, ### QN sections, ### Dimension Score)]

## Feature Necessity Audit
[from extras-generator output]

## Kill Criteria
[from extras-generator output]

## Pivot Directions
[from extras-generator output]

## Validation Playbook
[from extras-generator output — experiments targeting dimensions scored ≤3★]

Validate extras output:

• ## Elevator Pitch Test section exists with **Verdict:** field
• ## Kill Criteria section exists with >=3 numbered items
• ## Feature Necessity Audit section exists (content or skip notice)
• ## Pivot Directions section exists with >=2 named directions
• ## Validation Playbook section exists with >=2 numbered experiments (each with Do/Success/Fail/Timeline fields), or a skip notice if all dimensions scored >=4★

If validation fails: re-dispatch extras-generator once with a correction note. If still non-compliant, include with warning annotation.

Step 9.5: Save Report

If the evaluation target is a local project:

1. Check if docs/08-product-evaluation/ directory exists. If not, create it.
2. Save the report to docs/08-product-evaluation/{YYYY-MM-DD}-evaluate.md
3. Check for previous reports in the same directory (files matching *-evaluate.md). If found, append a ## Changes Since Last Evaluation section comparing dimension scores and highlighting movements >=1★.
   • Format: | Dimension | Previous | Current | Change | table
   • Only include dimensions with score movement >=1★
   • If this is the first report, omit this section

Step 10: Present Results

Display the assembled report.

Post-processing:

1. Highlight the Elevator Pitch result — if "Cannot articulate", call it out prominently at the top
2. Flag weak dimensions — any dimension scored <=2 stars gets an explicit warning
3. Verify Priority Actions section — confirm it lists all dimensions scored ≤3★ in ascending score order, each with its Next Action from the dimension-evaluator output. This replaces generic "next steps" with specific, dimension-derived actions.

Completion Criteria

• All 6 dimensions evaluated and validated
• Extras modules (Elevator Pitch, Kill Criteria, Feature Audit, Pivot Directions, Validation Playbook) generated
• Final report assembled with weighted total
• Report saved to docs/08-product-evaluation/ (if local project)