adev-assess

name

adev:assess

description

Assess codebase readiness for agentic development. Scans the codebase using static file inspection and outputs maturity scores across 8 structural dimensions (raw mode) or 11 dimensions (adev mode). Use when you want to evaluate how well-suited a codebase is for AI-assisted development, identify improvement areas, or track progress over time. In OpenCode, invoke with skill({ name: 'adev:assess' })

Codebase Readiness Assessment

Assess how well a codebase is prepared for agentic development. Runs static file analysis across multiple dimensions to produce a maturity score and actionable feedback.

Arguments

• No arguments: auto-detect mode based on presence of .context-index/
• --mode raw: assess only 8 structural dimensions (ignore adev-specific dimensions)
• --mode adev: assess all 11 dimensions (8 structural + 3 adev-specific)
• --output markdown (default): output as human-readable markdown scorecard
• --output json: output as machine-parseable JSON
• --target <path>: directory to assess (default: cwd)

Prerequisites

None — this skill operates entirely through static file inspection.

Load Skill Extensions

Load Skill Extensions: Load any skill extension instructions before proceeding:

adev skill-ext load --skill assess

If the output is not __NONE__, incorporate it as additional standing instructions that apply to this skill's entire execution. Frame it as: "The following skill extension instructions apply to this invocation (source: installed domain extensions and/or project-level overrides)." If the output is __NONE__, continue normally.

Assessment Dimensions

Structural Dimensions (8) — Used in both raw and adev modes

Dimension	Weight	What It Measures
Test Infrastructure	12.5%	Presence of test files, test framework, test patterns
Type Safety	12.5%	TypeScript/type annotations, strict mode, type coverage
Modularity	12.5%	File organization, clear boundaries, single responsibility
Naming	12.5%	Consistent naming conventions, descriptive names
Documentation	12.5%	README, comments, docstrings, API docs
Dependency Hygiene	12.5%	Package.json quality, lock files, dependency clarity
Build Configuration	12.5%	Build scripts, lint config, typecheck config
Spec Sources	12.5%	Architecture docs, specs, ADRs

Adev-Specific Dimensions (3) — Only used in adev mode

Dimension	Weight	What It Measures
Adev Context Index	33%	Presence and quality of .context-index/
Adev Skills	33%	Skill coverage and quality
Adev Hooks	33%	Hook configuration and coverage

Scoring Criteria

Test Infrastructure (0-100)

Score	Criteria
80-100	Test files present, test framework configured, >50% code coverage
50-79	Test files present, framework configured, some coverage
20-49	Some test files exist
0-19	No tests or minimal test files

Evidence: Glob for **/*.test.*, **/*.spec.*, **/tests/**, check package.json for test scripts

Type Safety (0-100)

Score	Criteria
80-100	TypeScript with strict mode, all files typed
50-79	TypeScript with strict mode, some files untyped
20-49	TypeScript or JSDoc types in some files
0-19	No type annotations

Evidence: Check for tsconfig.json, *.ts files, checkJs in jsconfig

Modularity (0-100)

Score	Criteria
80-100	Clear directory structure, single responsibility, small files (<300 lines)
50-79	Reasonably organized, some large files
20-49	Some organization, mixed patterns
0-19	Monolithic or disorganized

Evidence: Glob for directory structure, check file sizes

Naming (0-100)

Score	Criteria
80-100	Consistent naming across codebase
50-79	Mostly consistent with minor exceptions
20-49	Inconsistent naming patterns
0-19	Poor or unclear naming

Evidence: Grep for naming patterns, check file names

Documentation (0-100)

Score	Criteria
80-100	README, API docs, inline comments, good coverage
50-79	README present, some documentation
20-49	Minimal documentation
0-19	No documentation

Evidence: Check for README.md, docs/, src/**/*.md

Dependency Hygiene (0-100)

Score	Criteria
80-100	package.json clean, lock file present, reasonable dep count
50-79	package.json exists, some issues
20-49	package.json present but messy
0-19	No package.json or critical issues

Evidence: Read package.json, check for package-lock.json or yarn.lock

Build Configuration (0-100)

Score	Criteria
80-100	Build scripts, lint config, typecheck all present
50-79	Some build config present
20-49	Minimal build config
0-19	No build configuration

Evidence: Check for build scripts in package.json, .eslintrc, tsconfig

Spec Sources (0-100)

Score	Criteria
80-100	Architecture docs, ADRs, specs present
50-79	Some documentation exists
20-49	Minimal documentation
0-19	No architectural documentation

Evidence: Check for docs/, ADRs, spec files

Adev Context Index (0-100) — adev mode only

Score	Criteria
80-100	Complete .context-index/ with constitution, charters, specs, ADRs
50-79	Partial .context-index/ setup
20-49	Minimal .context-index/
0-19	.context-index/ missing or empty

Evidence: Check .context-index/ directory contents

Adev Skills (0-100) — adev mode only

Score	Criteria
80-100	Multiple well-structured skills present
50-79	Some skills defined
20-49	Minimal skills
0-19	No skills

Evidence: Check skills/ directory

Adev Hooks (0-100) — adev mode only

Score	Criteria
80-100	Multiple hooks configured and working
50-79	Some hooks present
20-49	Minimal hooks
0-19	No hooks

Evidence: Check hooks/ directory and hooks.json

Maturity Levels

Level	Score Range	Description
L1	0-20	Initial — minimal structure
L2	21-40	Developing — basic structure in place
L3	41-60	Defined — moderate preparedness
L4	61-80	Managed — well-prepared
L5	81-100	Optimized — excellent agentic readiness

Output Format

Persona adaptation: The formats below are defaults for the Developer persona. If a different persona is active, adapt the chat summary to its output rules. The scorecard file written to disk always uses the full format.

Markdown Scorecard

# Codebase Readiness Assessment

**Total Score:** 72/100 (L4 - Managed)
**Mode:** adev
**Assessed:** 2026-03-24T10:30:00Z

## Dimensions

| Dimension | Score | Indicator |
|-----------|-------|-----------|
| Test Infrastructure | 85 | 🟢 ████████░ |
| Type Safety | 78 | 🟢 ███████░░ |
| Modularity | 72 | 🟡 ██████░░░ |
| Naming | 65 | 🟡 ██████░░░ |
| Documentation | 55 | 🟡 █████░░░░ |
| Dependency Hygiene | 80 | 🟢 ███████░░ |
| Build Configuration | 75 | 🟢 ███████░░ |
| Spec Sources | 45 | 🟡 █████░░░░ |
| Adev Context Index | 90 | 🟢 █████████ |
| Adev Skills | 70 | 🟡 ███████░░ |
| Adev Hooks | 60 | 🟡 ██████░░░ |

JSON Output

{
  "version": "1.0.0",
  "timestamp": "2026-03-24T10:30:00Z",
  "mode": "adev",
  "totalScore": 72,
  "level": "L4",
  "dimensions": [
    {
      "name": "Test Infrastructure",
      "score": 85,
      "weight": 0.125,
      "evidence": ["Found test files", "Jest configured"]
    }
  ]
}

Process

1. Detect mode: Check for .context-index/ directory (or use explicit --mode)
2. Select dimensions: Load appropriate dimension set based on mode
3. Scan codebase: Use Glob/Grep to find evidence for each dimension
4. Score each dimension: Apply scoring criteria to produce 0-100 score
5. Calculate total: Weighted average of dimension scores
6. Determine level: Map total to L1-L5
7. Format output: Generate markdown or JSON per --output flag

Notes

• Uses only Node.js built-ins (fs, path) for file inspection
• No external commands are executed (no npm test, tsc, etc.)
• Evidence is collected for each dimension to explain scores
• Scores are deterministic and reproducible