fix(ci): harden smoke and benchmark command execution

[eduardbar]

Refs #42

Summary

• Make repository smoke timeouts configurable via CLI/env with validated precedence.
• Raise CI smoke execution to an explicit 120 seconds without changing the local default.
• Align the ESLint plugin dependency and lockfile with published Drift 1.6.x.
• Forward benchmark runner flags (--runs, --warmup, --base) without weakening unrelated Commander parsing.

Verification

• Focused smoke/package/CI tests
• Full suite on local runtime where feasible
• Build, runtime policy, docs drift, perf checks
• Root and plugin package dry-runs
• Full CLI command matrix; benchmark flags confirmed
• Self-scan performed without changing rule weights or exclusions

Known environment notes

• Supported-runtime authority is GitHub Actions Node 20/22; local runtime is Node 24.
• Smoke and perf variability are tracked against clean-master baselines.
• Existing user documentation/OpenSpec working-tree files are intentionally excluded.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
site	Ready	Preview	Aug 3, 2026 3:48pm

[github-actions]

drift trust

drift trust

• Trust Score: 55/100
• Merge Risk: HIGH
• Target: /home/runner/work/drift/drift

Diff signals

• Base ref: origin/master
• Diff status: NEUTRAL
• Score delta: +0
• Issues: +0 new / -0 resolved
• Trust adjustment: +0 penalty / -0 bonus (net +0)

Top reasons

• Error-level issues: 9 error issue(s) increase merge volatility. (impact 22)
• Highest-risk file: /home/runner/work/drift/drift/site/src/App.jsx has hotspot risk 70/100. (impact 11)
• Maintenance hotspots: Maintenance risk is MEDIUM (38/100). (impact 10)
• Drift score pressure: Repository drift score is 4/100. (impact 2)

Fix priorities

• feat(review): drift review command with PR comment output #1 magic-number (info, x65, effort: low) - Address this rule in the highest-scored files first.
• feat(score): repository quality score breakdown in scan/json #2 large-function (error, x7, effort: high) - Extract smaller functions to reduce review complexity.
• feat(ai): ai_likelihood and files_suspected in AI output #3 unused-export (warning, x10, effort: medium) - Address this rule in the highest-scored files first.
• feat(architecture): configurable architecture rules in drift.config.ts #4 large-file (error, x2, effort: high) - Split monolithic files by responsibility before merge.
• feat(map): drift map command to generate architecture.svg #5 no-return-type (info, x11, effort: low) - Address this rule in the highest-scored files first.

drift review

drift review

ℹ️ No drift changes detected against base ref.

• Base ref: origin/master
• Score delta: +0
• New issues: 0
• Resolved issues: 0

File breakdown

• No file-level deltas detected