Deterministic Stop-gate and code-graph context for AI coding agents: blocks only net-new findings and gives the loop a structural map of the codebase, locally, with no model in the gate.
Command
Per-finding suppression with a typed-category audit trail. The allowlist is dxkit's deliberate escape hatch for findings you can't fix today: false positives, intentional test-fixture patterns, real risks mitigated externally, or work you're deferring with an explicit deadline.
Command
Capture the per-finding identities of every issue dxkit currently surfaces in your repo, and write them to a JSON file under .dxkit/baselines/. The file becomes the "brownfield anchor" the guardrail check command diffs new scans against to decide what is a net-new regression vs. pre-existing debt.
Command
Bill of Materials. The unified package-centric view — every dependency with its license, vulnerabilities, reachability evidence, KEV/EPSS enrichment, composite Risk score, and an actionable upgrade plan.
Command
List and dry-run the custom checks configured for this repo — user-declared repo invariants (.dxkit/policy.json:checks) plus the pack-declared built-in lint gate.
Command
Compute a deterministic configuration plan for this repo, and optionally apply it. Each capability in the command registry declares a pure planConfig probe that derives its recommended settings from observable repo facts, so the same repo yields the same plan on every run and in every environment. There is no judgment…
Command
Hand a coding agent (or yourself) a slim, token-budgeted structural slice for a query — the relevant symbols, where they live, what calls them, and the module they belong to — read from the code graph at .dxkit/reports/graph.json. The point is to navigate by graph instead of repeated whole-file reads: the same…
Command
Run each active language pack's test command with coverage instrumentation and write the materialized coverage artifact. Used to upgrade test-gaps + health.Testing from filename-match heuristic to real line-coverage truth.
Command
Single-page HTML view assembling every report that's been generated. Doesn't run analysis — reads what's already in .dxkit/reports/.
Command
Deterministic dependency security bumps: turn the fixable subset of your dependency vulnerabilities into concrete, verified upgrades — planned from the scanners' own fix versions, applied with your repo's own package manager, verified by the correctness floor and the guardrail before anything lands. No LLM anywhere in…
Command
A shareable snapshot of what dxkit sees in a repository: the whole code graph (every function, what it calls, and what calls it), joined to the HTTP contract when one exists (routes served, calls made, and how they bind). It reads the same deterministic analysis the gate uses and adds no new heuristics.
Command
Developer activity report. Per-author commit + line stats, per-week velocity, and a hot-files ranking.
Command
Diagnose missing tools, version mismatches, environment misconfig. The first stop when something doesn't work.
Command
A zero-write trial. evaluate replays your repo's recent landings (the last few merged changes, or an explicit before/after ref pair) through the same deterministic guardrail gate the hooks and CI would run, and reports what the gate would have blocked plus what enabling dxkit costs (measured gate latency, interruption…
Command
Ask the codebase structural questions — what it does, where a feature lives, which files everything depends on. Every subcommand reads the same deterministic code graph at .dxkit/reports/graph.json (symbols, call edges, and Louvain-clustered modules, extracted by graphify). It doesn't run analysis; it queries the…
Command
Diff a current scan against a committed baseline and classify each finding as added / relocated / toolingdrift / configdrift / persisted / removed / fixed. Block on net-new regressions per the brownfield policy; exit non-zero so a pre-commit hook, pre-push hook, or CI workflow can stop the commit/push/merge.
Command
The flagship report. Produces a 6-dimension 0-100 health score with per-dimension metrics and ranked remediation actions.
Command
Install the dxkit agent DX layer in a repository: AGENTS.md + CLAUDE.md shim + .claude/skills/dxkit-/ (the lifecycle skills).
Command
Open a pre-filled GitHub Issue against vyuh-labs/dxkit in your default browser. Use it when you want to report a false positive, a missing finding, a dxkit bug, a feature request, or a docs gap.
Command
The operability answer to "which dxkit jobs run here, when, and did they work?". One read-only view over the installed dxkit workflows: each trigger, the actual cron parsed from the workflow file, the computed next run (UTC), the last run's outcome when the gh CLI is available, and the run-it-now command for anything…
Command
The loop pack is dxkit's deterministic preflight/postflight layer for autonomous coding loops — a Claude Code session that keeps working until it decides to stop. It stops a loop from declaring "done" while it has introduced net-new findings, by re-running the guardrail on every Stop and feeding any net-new findings…
Command
Serve dxkit's read-only repo tools over the Model Context Protocol, so any MCP-capable coding agent (Claude Code, Cursor, VS Code) can call them natively while working in your repo.
Command
Code quality + slop detection. Lint summary, duplicate code (jscpd), structural metrics (graphify), and a "slop score" composite.
Command
Close the debt the deterministic lanes cannot version-bump away: the grandfathered broken build, advisories with no direct upgrade path, the lint backlog, missing tests, missing docs. Each run plans that debt as finite work orders and works them in two tiers: deterministic recipes execute first at $0 (lockfile resync…
Command
Run every analyzer + the dashboard in one shot. The full audit.