mock-production-gap
Mock-production gap detector: compare test mocks against real implementations, LLM judges dangerous divergences. Read-only. Trigger: /mock-gap
SKILL.mdskills/runtime/mock-production-gap/SKILL.md
GitHub/mock-gap
Tests pass with mocks; production breaks with real dependencies. This skill finds mock divergences before they ship.
What this is for
- Mock return values missing fields that production code uses
- Mock error types that don't match real error types
- Mock function signatures that differ from real implementations
- Read-only skill. No test changes.
PROTECTION RULE - never ~/.claude/
Read-only skill. Guard required if write mode added later.
What You Must Do When Invoked
During analysis, assign a confidence level to each finding: proven (confirmed by evidence), likely (strong signal, needs review), or suspected (weak signal).
If /mock-gap -help or /mock-gap -h (without further arguments)
is invoked: output the ## Usage section unchanged and stop.
Otherwise follow these steps in order, skipping none.
Step 1 - Help check
If invoked with -help or -h, output the ## Usage section unchanged and stop.
Step 2 - Clarify target
Clarify -ProjectDir. Get confirmation.
Step 3 - Scan
& "<SKILL_DIR>/scripts/mock-scan.ps1" -ProjectDir "<path>"
Step 4 - Classification
Read each mock divergence:
- Dangerous: divergent field is used in production code path (test doesn't test real behavior)
- Noisy: divergent field exists but isn't used by calling code
- Benign: mock intentionally simplified for unrelated tests
- Outdated: real API changed, mock not updated
Step 5 - Write report
File mock-gap-report.md in current working directory:
- Summary - divergences by severity.
- Divergence table - dangerous first. Per divergence: test file, mocked module, divergent field, mock value, real value, severity, recommendation.
- Open questions.
Step 6 - Summarize
State report path, highlight mocks that silently test wrong behavior.
Usage
/mock-gap # interactive
/mock-gap <dir> # scan project
/mock-gap -help