test-gap-analysis · diff
git:20260914.ba5cf40 to git:20260914.45aa88f
17 added, 1 removed. Audit A to A.
---
name: test-gap-analysis
description: Use this skill when you need to identify missing or weak test obligations from requirements, risks, changes, defects, and test evidence; triggers include test gap analysis.
---
# Test Gap Analysis
Find test obligations that are not sufficiently protected by test intent or execution evidence across requirements, risks, changes, defects, and test assets. Produce actionable `TG-##` gaps. This is not a replacement for a traceability matrix, test-case review, or test execution.
## When to Use
- Use it to determine which requirements, risks, behaviors, or failure modes lack a test intent or evidence.
- Use it to find orphan tests, stale evidence, unverified execution, low-value duplicates, and high-risk uncovered obligations.
- Use it when inputs are incomplete but a bounded gap register and evidence questions are still useful.
Do not use it only to write test cases, build a complete `RT-##`/`TC-##` matrix, execute tests, or accept residual risk for a Human.
+ ## Output Format Options
+
+ - Use Markdown by default; when a table, CSV, or JSON is requested, preserve the same evidence, status, impact, owner, and validation fields.
+ - Do not present a structured format or static inventory as execution, pass, approval, or release evidence.
+
+ ## How to Use
+
+ 1. Read this Skill's primary prompt and provide the objective, scope, material, environment, and available evidence.
+ 2. Follow the prompt's input audit and output contract; deliver a bounded first pass when information is incomplete.
+ 3. Retain source, evidence status, impact, owner role, close condition, and validation method for every finding.
+
## Workflow
1. Read and follow `prompts/test-gap-analysis.md`, beginning with the six-part input audit.
2. Compare test obligations with test assets in both directions; never infer coverage from filenames or titles alone.
3. Use `TG-##` to distinguish missing mappings, orphan tests, unverified execution, stale evidence, uncovered risks, and low-value duplicates.
4. Preserve source, evidence state, impact/priority, proposed test intent, owner role, and close condition for every gap.
5. When evidence is missing, return a bounded result and mark `unassessed`, `unverified`, or `blocked` boundaries.
## Core Constraints
- `TG-##` records a gap and an action, not a coverage claim, pass, release, or risk-acceptance evidence.
- File presence, similar test names, report summaries, or static configuration cannot alone prove execution or coverage.
- Every material conclusion needs a source and minimum evidence; inferences must state assumptions and validation.
- Never invent requirements, priorities, test results, defect states, owners, or closure facts.
- Do not duplicate complete traceability analysis, test-case authoring, or executable-set selection here.
- ## Progressive Disclosure
+ ## Reference Files
- Always read `prompts/test-gap-analysis.md` before producing an analysis.
- For regression, read `evals/eval.yaml` and matching `evals/cases/`; Eval configuration does not prove project results.
- For trigger checks, use `evals/trigger-prompts.csv` and `evals/local-rules.json`; missing `skill.selection` evidence is `BLOCKED`.
+
+ ## Best Practices
+
+ - Prioritize high-impact gaps with a verifiable next action, using the smallest useful experiment or evidence request.
+ - Separate facts, evidence-backed inferences, recommendations, and Human decisions; never upgrade an assumption into a conclusion.
## Pre-delivery Check
- [ ] Recorded known facts, missing information, conflicts, stale information, out-of-scope items, and assumptions.
- [ ] Each `TG-##` has an obligation, source, gap type, evidence, impact/priority, and closure action.
- [ ] Static presence, inference, unverified execution, and actual execution evidence remain separate.
- [ ] High-risk gaps have an owner role, smallest evidence action, and validation method.
- [ ] The gap register is not presented as full coverage, pass, release, or Human risk acceptance.
## Common Pitfalls
- Treating a test file as proof that a requirement is covered.
- Saying “coverage is insufficient” without identifying the missing obligation and evidence.
- Marking every gap as highest priority or turning a suggested test into an execution result.
- Hiding the actual gap judgment inside a full traceability matrix or a test-case dump.