Skip to content
Back to skills

Review Codebase Analysis

ASecurity

Review a codebase analysis for coverage of relevant components, accuracy of behavior claims, rigor of changeability assessments, and completeness of migration and impact analysis.

  • 10 stars
  • 0 votes
  • 0 copies
  • 0 views
  • Added October 6, 2026
developmentgonodeapi

Works with

  • cli
  • api

Security analysis

A100/100

Scanned October 6, 2026

npx -y skills add tomzx/agents --skill review-codebase-analysis --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Review Codebase Analysis?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Review Codebase Analysis
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/tomzx-review-codebase-analysis/badge)](https://www.skillsdirectory.com/skills/tomzx-review-codebase-analysis)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: review-codebase-analysis
description: Review a codebase analysis for coverage of relevant components, accuracy of behavior claims, rigor of changeability assessments, and completeness of migration and impact analysis.
---

# Review Codebase Analysis

Audits a codebase analysis and reports findings across five categories: coverage, accuracy, changeability rigor, and impact and migration completeness.
The review verifies that the analysis describes the actual code (not assumptions) and that every change disposition is justified and its risk addressed.

## Prerequisites

- Apply the shared SDLC conventions in `skills/sdlc/references/shared.md`.
- If no argument is provided, locate the feature directory under `.sdlc/features/` whose frontmatter `issue` field references `$ISSUE_NUMBER`.
- `.sdlc/features/N-<slug>/codebase-analysis.md`, or an analysis document provided in context or as a file path
- `.sdlc/features/N-<slug>/requirements.md` (optional, improves coverage analysis)

## Steps

1. Read the analysis from `.sdlc/features/N-<slug>/codebase-analysis.md` if present, otherwise from context or as a file path.
2. Cross-reference against the requirements document if available: every requirement that implies a code change should map to an analyzed component.
3. Spot-check the analysis against the actual codebase to confirm behavior claims and paths.
4. Identify issues in each of the categories below.
5. Report findings. Omit any category that has no findings.
6. Write the findings to `.sdlc/features/N-<slug>/review-codebase-analysis.md` with frontmatter `artifact: codebase-analysis`, `verdict` (`approved` if there are no blocking findings, `changes-requested` if the author must address findings, `rejected` for a fundamental flaw), and `reviewed_at: <ISO date>`, and the findings as the body, per `skills/sdlc/references/shared.md`. Record any unresolved open questions in the findings body. For any question that carries meaningful risk, also invoke `/create-assumption` to record it formally. For a chosen change disposition with lasting consequences (e.g. replace vs. extend), invoke `/create-decision`.

## Review Checklist

### Coverage
- Does the analysis address every requirement that implies a change to existing code, or are touched components missing?
- Are the search entry points recorded (queries, paths) so the scope is auditable?
- For a greenfield verdict, is the integration boundary (where it attaches to existing systems) actually empty, or was existing code overlooked?

### Accuracy
- Are component responsibilities and paths correct and current (verify against the codebase, not the prose)?
- Are behavior claims (ordering guarantees, error handling, side effects) backed by the source rather than assumed?
- Is any stale or recently changed code treated as current?

### Changeability Rigor
- Is each component assigned exactly one disposition (reuse, extend, refactor, replace)?
- Is every disposition justified by a rationale tied to the requirements and the coupling map, not by preference?
- Is the risk of each disposition stated, and is the risk driver concrete?
- Are the "must not change" constraints (public API, data formats, behavior contracts) explicit for extend/refactor/replace?

### Impact and Migration
- For every refactor or replace disposition, is the blast radius mapped (what else depends on the changed part)?
- Is there a migration path from current to target behavior, with backward compatibility and rollout strategy?
- Are de-risking measures considered (feature flag, dual-run, shadow comparison) for high-risk changes?
- If the section is omitted, is that justified (i.e. only reuse/extend dispositions, or greenfield)?

### Coupling Awareness
- Is the dependency map rendered as a Mermaid flowchart whose node classes match each component's change disposition?
- Does every edge in the diagram correspond to a real dependency (spot-check against the source), and is any real dependency missing from the diagram?
- Are dependencies between relevant components and external systems mapped, including shared state and synchronous vs. asynchronous boundaries?
- Does the changeability assessment account for the coupling, or does it ignore downstream effects?

## Output Format

```markdown
## Coverage

<Findings or "No issues found.">

## Accuracy

<Findings or "No issues found.">

## Changeability Rigor

<Findings or "No issues found.">

## Impact and Migration

<Findings or "No issues found.">

## Coupling Awareness

<Findings or "No issues found.">
```

## Outcome

If `$OUTCOME_YAML` is set, emit your verdict there per `skills/sdlc/references/shared.md`:

| Verdict | When |
|---|---|
| `approved` | No blocking findings; the subject passes review |
| `changes-requested` | Findings the author must address before it passes |
| `rejected` | Fundamental flaw requiring rework or stopping |

In the same emission, list the findings file under `artifacts:` (`.sdlc/features/N-<slug>/review-codebase-analysis.md`).

## Example Usage

**Scenario 1: A touched component was missed**
The requirements imply changes to a shared cache layer, but the analysis only covers the reconciliation loop.
Report under Coverage.

**Scenario 2: Behavior claim contradicts the code**
The analysis claims a consumer reads asynchronously, but the source shows a synchronous call, which changes the blast radius.
Report under Accuracy.

**Scenario 3: Replace disposition without a migration path**
The analysis recommends replacing the polling loop with an event-driven consumer but gives no rollout, backward-compatibility, or de-risking plan.
Report under Impact and Migration.

**Scenario 4: Risk stated without a driver**
A component is marked "Replace, High risk" with no explanation of what drives the risk.
Report under Changeability Rigor.

## Next Step

Once the findings verdict is `approved`, continue with `/create-feasibility`, which consumes this analysis to judge viability and cost.

## Useful Commands Reference

| Command | Description |
|---|---|
| `read` / `grep` | Verify component paths and behavior claims against the actual source |
| `mmdc -i <diagram.mmd>` or `npx -y @mermaid-js/mermaid-cli` | Best-effort Mermaid render check for the coupling flowchart |
| `/create-assumption` | Formalize an unresolved open question that carries risk |
| `/create-decision` | Record a change disposition with lasting consequences |

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…