Generate and manage project quality evaluation reports from feature perspective. Evaluates requirements, features, test coverage, and code alignment status with gap analysis. Generates consistent markdown reports to x-ipe-docs/quality-evaluation folder. Triggers on requests like "evaluate project quality", "generate quality report", "assess code alignment".
Standardmäßig ist der Prompt ausgewählt, der zuerst die Quelle prüft. Sie können zu einem direkten Befehl wechseln oder eine lokale Kopie herunterladen.
Quelldateien prüfen
Lesen Sie SKILL.md und alle von SkillsMP angezeigten Begleitdateien, bevor Sie sich für eine Installation entscheiden.
Mit Codex oder Claude installieren Kopieren Sie diesen Prompt, fügen Sie ihn in Codex, Claude oder einen anderen Assistant ein und lassen Sie die Skill-Seite prüfen und installieren.
Ein direkter Befehl überspringt den Prüf-Prompt. Prüfen Sie die Quelle, bevor Sie ihn ausführen.
Generate and manage project quality evaluation reports from feature perspective. Evaluates requirements, features, test coverage, and code alignment status with gap analysis. Generates consistent markdown reports to x-ipe-docs/quality-evaluation folder. Triggers on requests like "evaluate project quality", "generate quality report", "assess code alignment".
Project Quality Board Management
Purpose
AI Agents follow this skill to generate and manage project-wide quality evaluation reports:
Generate quality evaluation reports across all features
Update existing reports with re-evaluated features
Query quality status for specific features
Compare quality between evaluation snapshots
Important Notes
BLOCKING: Learn x-ipe-workflow-task-execution skill before executing this skill. If Agent does not have skill capability, go to .github/skills/ folder to learn skills. SKILL.md is the entry point for each skill.
CRITICAL: Learn x-ipe-tool-refactoring-analysis skill to understand refactoring_suggestion and refactoring_principle data models for integration.
About
This skill evaluates the project from a feature perspective, analyzing four primary dimensions plus tracing and security coverage.
Key Concepts:
Quality Evaluation -- Scored assessment (1-10) of a feature across six dimensions: requirements, specification, test coverage, code alignment, tracing, security
Health Status -- Overall project health: healthy (8-10), attention_needed (5-7), (1-4)
<definition_of_ready><checkpointrequired="true"><name>Operation identified</name><verification>Caller specifies generate, update, query, or compare</verification></checkpoint><checkpointrequired="true"><name>Feature discovery path exists</name><verification>x-ipe-docs/requirements/FEATURE-* directories are scannable</verification></checkpoint><checkpointrequired="true"><name>Evaluation references available</name><verification>references/evaluation-principles.md and references/evaluation-procedures.md exist</verification></checkpoint></definition_of_ready>
Operations
Operation: Generate Quality Report
When: Need to evaluate project quality from scratch or on schedule
<operationname="generate"><action>
1. DETERMINE scope (default: all features; filter by feature_ids or module_paths if provided)
2. DISCOVER features by scanning x-ipe-docs/requirements/FEATURE-* directories
3. FOR EACH feature:
a. EVALUATE requirements alignment (see references/evaluation-procedures.md)
b. EVALUATE specification alignment
c. EVALUATE test coverage
d. EVALUATE code alignment
e. EVALUATE tracing coverage
f. EVALUATE security
g. COLLECT violations per category
h. CALCULATE feature score using dimension weights (req:0.20, spec:0.20, test:0.20, code:0.20, tracing:0.10, security:0.10)
4. AGGREGATE results: calculate overall_score (average of non-planned features), determine health_status, collect priority_gaps, generate recommendations
5. GENERATE report using template at references/quality-report-template.md
6. SAVE report with versioning (see Versioning Workflow below)
7. SELF-REVIEW: read report completely, check for missing content, inconsistencies, features mentioned but not evaluated, gaps without severity, recommendations not matching gaps; fix any issues found
</action><constraints>
- BLOCKING: Report MUST follow structure rules (see Report Structure Rules below)
- CRITICAL: Planned features are excluded from score calculation
- CRITICAL: Report header must include Project Version (from pyproject.toml) and Evaluated Date
</constraints><output>Evaluation summary with overall_score, health_status, feature_count, gap_counts</output></operation>
Operation: Update Existing Report
When: Need to re-evaluate specific features without full project scan
<operationname="update"><action>
1. LOAD latest report from x-ipe-docs/quality-evaluation/project-quality-evaluation.md
2. FOR EACH feature_id in scope:
a. RE-EVALUATE all six dimensions
b. UPDATE feature entry in report
3. RE-CALCULATE aggregates (overall_score, health_status)
4. UPDATE priority_gaps and recommendations
5. SAVE using versioning workflow
</action><output>Updated evaluation summary with changed features highlighted</output></operation>
Operation: Query Quality Status
When: Need to check quality for specific features without generating a new report
<operationname="query"><action>
1. LOAD latest report from x-ipe-docs/quality-evaluation/project-quality-evaluation.md
2. FILTER features by provided criteria (feature_ids, status, score range)
3. RETURN filtered evaluation data
</action><output>Matching feature evaluations or empty list</output></operation>
Operation: Compare Evaluations
When: Need to track quality changes over time between two report versions
<operationname="compare"><action>
1. LOAD both evaluation reports (latest vs previous, or v{A} vs v{B})
2. FOR EACH feature present in both:
a. COMPARE status changes
b. CALCULATE score deltas per dimension
c. IDENTIFY new and resolved gaps
3. GENERATE comparison summary: improved features, degraded features, new gaps introduced, gaps resolved
</action><output>Comparison data with score deltas and gap changes per feature</output></operation>
Versioning Workflow
Used by generate and update operations when saving reports.
<operationname="version_report"><action>
1. ENSURE x-ipe-docs/quality-evaluation/ folder exists (create if not)
2. IF project-quality-evaluation.md exists:
a. SCAN for existing project-quality-evaluation-v*.md files
b. FIND max version number N
c. RENAME current file to project-quality-evaluation-v{N+1}.md
d. IF more than 4 versioned files: DELETE oldest (lowest version number)
3. SAVE new report as project-quality-evaluation.md
4. VERIFY: at most 1 current + 4 historical files
</action><constraints>
- BLOCKING: Never exceed 5 total report files
- CRITICAL: Version number = max(existing versions) + 1
</constraints><output>Report file path and version number</output></operation>
Report Structure Rules
<constraintsname="report_structure_rules"><ruleid="1">Evaluation Principles section MUST come BEFORE Feature-by-Feature Evaluation; MUST only explain principles and thresholds; MUST NOT contain evaluation results</rule><ruleid="2">Violation Details MUST be organized by feature with 6 subsections each: Requirements, Specification, Test Coverage, Code Alignment, Tracing Coverage, Security; ONLY show features with violations; show "No violations" for clean categories</rule><ruleid="3">Separation of concerns: Principles = WHAT and HOW (thresholds); Violations = RESULTS per feature</rule><ruleid="4">Tracing: threshold >=90% decorator coverage; sensitive params (password, token, secret, key) require redact=[]; levels: API=INFO, Business=INFO, Utilities=DEBUG</rule><ruleid="5">Security: input validation on endpoints, no hardcoded secrets, auth on protected routes, injection prevention, secure sensitive data handling</rule></constraints>
<definition_of_done><checkpointrequired="true"><name>All features in scope evaluated</name><verification>Every feature in scope has scores for all six dimensions</verification></checkpoint><checkpointrequired="true"><name>Gaps identified and prioritized</name><verification>All gaps have severity assigned and are collected in priority_gaps</verification></checkpoint><checkpointrequired="true"><name>Overall score calculated</name><verification>overall_score and health_status are set based on feature scores</verification></checkpoint><checkpointrequired="true"><name>Report saved with versioning</name><verification>Report exists at correct path; historical versions preserved per versioning rules</verification></checkpoint><checkpointrequired="true"><name>Self-review completed</name><verification>Report read completely; no missing content, inconsistencies, or unassigned severities</verification></checkpoint></definition_of_done>
Error Handling
Error
Cause
Resolution
NO_FEATURES_FOUND
No FEATURE-* directories in requirements
Return empty evaluation; suggest running feature-breakdown first
REPORT_NOT_FOUND
No existing report for update/query/compare
Run generate operation first
VERSION_CONFLICT
Version number collision during save
Re-scan for max version and retry
EVALUATION_INCOMPLETE
A dimension could not be evaluated (missing docs/code)
Mark dimension as planned; exclude from scoring
SCOPE_INVALID
Provided feature_ids or module_paths not found
Return error with list of valid feature_ids
Templates
File
Purpose
references/quality-report-template.md
Report template with sections and placeholder fields
references/evaluation-principles.md
Evaluation principles, thresholds, and score formulas
references/evaluation-procedures.md
Step-by-step procedures for each evaluation dimension
input:operation:"generate"scope:type:"feature_ids"filter: ["FEATURE-001", "FEATURE-003"]
# After code-refactoring-stage, re-evaluate affected features and compare with baseline
