Product Behavior Auditor (L3 Worker) Skill

Paths: File paths (references/, ../ln-*) are relative to this skill directory.

Product Behavior Auditor (L3 Worker)

Type: L3 Worker

Specialized worker auditing whether tests prove product behavior instead of platform behavior.

Purpose & Scope

Audit Product Behavior Focus (Category 1: High Priority)
Detect tests validating language, framework, library, generated, or default platform behavior instead of local product logic
Emit DELETE_NON_PRODUCT_TEST or REWRITE_TO_PRODUCT_BEHAVIOR
Calculate compliance score (X/10)

Inputs

MANDATORY READ: Load references/audit_worker_core_contract.md.

Receives contextStore with: tech_stack, testFilesMetadata, codebase_root, output_dir.

Workflow

Detection policy: use two-layer detection (candidate scan, then context verification); load references/two_layer_detection.md only when the verification method is ambiguous.

Parse Context: Extract tech stack, framework detection patterns, test file list, output_dir from contextStore
Scan Codebase (Layer 1): Scan test files for framework/library tests (see Audit Rules below) 2b) Context Analysis (Layer 2 -- MANDATORY): For each candidate, read test code and ask:
- Does this test custom code that wraps a framework primitive (e.g., custom hook using useState)? -> KEEP (testing integration, not framework)
- Does this test ONLY call language/framework/library API with no custom product logic? -> DELETE_NON_PRODUCT_TEST
- Can the test be rewritten to assert local rules, mappings, policies, or error behavior? -> REWRITE_TO_PRODUCT_BEHAVIOR
- Is this a test helper/utility that imports libraries for mocking setup? -> skip (not a test of framework behavior)
Collect Findings: Record each violation with severity, location (file:line), effort estimate (S/M/L), recommendation
Calculate Score: Count violations by severity, calculate compliance score (X/10)
Write Report: Build full markdown report in memory per references/templates/audit_worker_report_template.md, write to {output_dir}/ln-631--global.md in single Write call
Return Summary: Return minimal summary to coordinator (see Output Format)

Audit Rules

1. Framework Tests Detection

What: Tests validating framework behavior (Express, Fastify, Koa) instead of OUR business logic

Detection Patterns:

(express|fastify|koa).(use|get|post|put|delete|patch)
Test names: "middleware is called", "route handler works", "Express app listens"

Severity: MEDIUM

Recommendation: DELETE_NON_PRODUCT_TEST when the test only validates framework behavior. Use REWRITE_TO_PRODUCT_BEHAVIOR when a focused assertion can prove local integration logic.

Effort: S (delete test file or test block)

2. ORM/Database Library Tests

What: Tests validating Prisma/Mongoose/Sequelize/TypeORM behavior

Detection Patterns:

(prisma|mongoose|sequelize|typeorm).(find|findMany|create|update|delete|upsert)
Test names: "Prisma findMany returns array", "Mongoose save works"

Severity: MEDIUM

Recommendation: DELETE_NON_PRODUCT_TEST when the test only validates ORM behavior. Use REWRITE_TO_PRODUCT_BEHAVIOR for repository policies, query composition, error mapping, or transaction rules.

Effort: S

3. Crypto/Hashing Library Tests

What: Tests validating bcrypt/argon2 hashing behavior

Detection Patterns:

(bcrypt|argon2).(hash|compare|verify|hashSync)
Test names: "bcrypt hashes password", "argon2 compares correctly"

Severity: MEDIUM

Recommendation: DELETE_NON_PRODUCT_TEST when the test only validates library behavior. Use REWRITE_TO_PRODUCT_BEHAVIOR for password policy, credential lifecycle, or wrapper error handling.

Effort: S

4. JWT/Token Library Tests

What: Tests validating JWT signing/verification

Detection Patterns:

(jwt|jsonwebtoken).(sign|verify|decode)
Test names: "JWT signs token", "JWT verifies signature"

Severity: MEDIUM

Recommendation: DELETE_NON_PRODUCT_TEST when the test only validates JWT library behavior. Use REWRITE_TO_PRODUCT_BEHAVIOR for token claims, expiry policy, roles, or auth flow.

Effort: S

5. HTTP Client Library Tests

What: Tests validating axios/fetch/got behavior

Detection Patterns:

(axios|fetch|got|request).(get|post|put|delete|patch)
Test names: "axios makes GET request", "fetch returns data"

Severity: MEDIUM

Recommendation: DELETE_NON_PRODUCT_TEST when the test only validates HTTP client behavior. Use REWRITE_TO_PRODUCT_BEHAVIOR for retry policy, timeout policy, request shaping, or error mapping.

Effort: S

6. React Hooks/Framework Tests

What: Tests validating React hooks behavior (useState, useEffect, etc.)

Detection Patterns:

(useState|useEffect|useContext|useReducer|useMemo|useCallback)
Test names: "useState updates state", "useEffect runs on mount"

Severity: LOW (acceptable if testing OUR custom hook logic)

Recommendation: DELETE_NON_PRODUCT_TEST if testing framework behavior. Use REWRITE_TO_PRODUCT_BEHAVIOR if a custom hook or component policy can be asserted through product-visible behavior.

Effort: S-M

Scoring Algorithm

MANDATORY READ: Load references/audit_scoring.md.

Output Format

MANDATORY READ: Load references/templates/audit_worker_report_template.md.

Write JSON summary per references/audit_summary_contract.md. In managed mode the caller passes both runId and summaryArtifactPath; in standalone mode the worker generates its own run-scoped artifact path per shared contract.

Write report to {output_dir}/ln-631--global.md with category: "Product Behavior Focus" and checks: framework_tests, orm_tests, crypto_tests, jwt_tests, http_client_tests, react_hooks_tests. Findings must include action as DELETE_NON_PRODUCT_TEST or REWRITE_TO_PRODUCT_BEHAVIOR.

Return summary per references/audit_summary_contract.md.

When summaryArtifactPath is absent, write the standalone runtime summary under .hex-skills/runtime-artifacts/runs/{run_id}/evaluation-worker/{worker}--{identifier}.json and optionally echo the same summary in structured output.

Report written: .hex-skills/runtime-artifacts/runs/{run_id}/audit-report/ln-631--global.md
Score: X.X/10 | Issues: N (C:N H:N M:N L:N)

Critical Rules

Apply the already-loaded references/audit_worker_core_contract.md.

Do not auto-fix: Report only
Unique angle: Only decide whether the test proves local product behavior. Do not score portfolio value, E2E coverage, isolation, oracle strength, or structure.
Framework-specific patterns: Match detection patterns to project's actual tech stack
Effort realism: S = <1h, M = 1-4h, L = >4h
Context-aware: Custom wrappers around libraries (e.g., custom hook using useState) are OUR code -- do not flag
Exclude test helpers: Do not flag shared test utilities that import libraries for mocking setup

Definition of Done

Apply the already-loaded references/audit_worker_core_contract.md.

[ ] contextStore parsed successfully (including output_dir)
[ ] All 6 checks completed (framework, ORM, crypto, JWT, HTTP client, React hooks)
[ ] Findings collected with severity, location, effort, recommendation, and action
[ ] Score calculated using penalty algorithm
[ ] Report written to {output_dir}/ln-631--global.md (atomic single Write call)
[ ] Summary written per contract

Reference Files

Audit output schema: references/audit_output_schema.md

Version: 3.0.0 Last Updated: 2025-12-23

Agent Skills: Product Behavior Auditor (L3 Worker)

Install this agent skill to your local

Skill Files