Verification

This guide covers the lexicon verify command — the core of the verification pipeline. Verification runs gates, computes scores, checks coverage, and writes audit records.

Running Verification

lexicon verify

This executes the full pipeline:

1. Loads gate definitions from specs/gates.toml
2. Runs each gate command as a subprocess
3. Loads the scoring model from specs/scoring/model.toml
4. Maps gate results to scoring dimensions
5. Computes the weighted score and verdict
6. Writes an audit record to .lexicon/audit/
7. Prints a summary with gate results and score breakdown

Understanding the Output

Gate Results

Each gate reports one of:

• Pass — command exited with status 0
• Fail — command exited with non-zero status
• Skip — gate was in the skip list
• Error — command could not be executed

Score and Verdict

The verdict follows this logic:

1. If any required dimension failed → Fail
2. If total score >= pass threshold (default 80%) → Pass
3. If total score >= warn threshold (default 60%) → Warn
4. Otherwise → Fail

Score Breakdown

The lexicon verify output includes a detailed breakdown showing each dimension’s weight, value, and contribution to the total score.

Skipping Gates

You can skip specific gates:

lexicon verify --skip doc-tests

Only gates with allow_skip = true may be skipped. Required gates with allow_skip = false cannot be skipped — the policy validator rejects the request.

CI Integration

Add verification to your CI pipeline:

- name: Run lexicon verify
  run: lexicon verify

See the CI Integration guide for a complete GitHub Actions example.

Audit Trail

Every verification run is recorded as an audit entry in .lexicon/audit/. This provides a history of project health over time and ensures all verification results are attributable and reviewable.