Skip to content
Back to skills

Lev Sdlc

ASecurity

Thin skill documenting the `lev sdlc *` command surface, SDLC plugin profiles, and related work/review loop commands in Leviathan.

  • 22 stars
  • 0 votes
  • 0 copies
  • 3 views
  • Added September 20, 2026
ai-agentsbashgitdocumentation

Works with

  • claude code
  • cli

Security analysis

A100/100

Scanned September 20, 2026

npx -y skills add lev-os/agents --skill lev-sdlc --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Lev Sdlc?

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

Security grade badge for Lev Sdlc
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/lev-os-lev-sdlc/badge)](https://www.skillsdirectory.com/skills/lev-os-lev-sdlc)

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: lev-sdlc
description: Thin skill documenting the `lev sdlc *` command surface, SDLC plugin profiles, and related work/review loop commands in Leviathan.
---

# lev-sdlc — SDLC Poly Surface Reference

> Thin skill documenting all `lev sdlc *` commands and the SDLC plugin surface.
> Source: `plugins/core-sdlc/` in the Leviathan monorepo.

## Command Surface

### `lev sdlc exec`

Execute a chore/plan entity through the SDLC pipeline.

```bash
lev sdlc exec [entity] [--dry-run] [--json]
```

**Profile:** `sdlc.flowmind.exec`
**What it does:** Runs exec (opus-4-6) + validate (codex/gpt-5.4) phases on an entity.
**Source:** `plugins/core-sdlc/src/commands/sdlc.ts`

### `lev sdlc deepen`

Fan-out plan decomposition into parallel deep-research subagents.

```bash
lev sdlc deepen [entity] [--json]
```

**Profile:** `sdlc.flowmind.deepen`
**Source:** `plugins/core-sdlc/src/commands/sdlc.ts`

### `lev sdlc hygiene`

Periodic 3-axis scan: handoffs + chore-code alignment + spec update proposals.

```bash
lev sdlc hygiene [--json]
```

**Profile:** `sdlc.flowmind.hygiene`
**Source:** `plugins/core-sdlc/src/commands/sdlc.ts`

### `lev sdlc autodev`

Heartbeat-driven autonomous development loop.

```bash
lev sdlc autodev [--interval=5m] [--max-ticks=10] [--budget=50k] [--json]
```

Routes to `lev loop autodev`. See `/autodev-lev` skill for full documentation.
**Source:** `plugins/core-sdlc/src/commands/autodev.ts`

### `lev work`

SDLC work lifecycle router with stage-based dispatch.

```bash
lev work [query] [--stage=auto|find|gather|review|contract] [--dry-run] [--json] [--epic=ID]
```

**Stages:**
- `auto` — Classify intent and route automatically
- `find` — Discover entities and prior art
- `gather` — Collect context and evidence
- `review` — Review and validate entities
- `contract` — Generate/update BDD contracts

**Source:** `plugins/core-sdlc/src/commands/work.ts`

### `lev review`

Review queue for entities pending approval.

```bash
lev review [--limit=N] [--json]
lev review approve <id>
lev review reject <id>
lev review defer <id>
```

**Config:** `sdlc.gates.review_labels`, `sdlc.review.default_limit`
**Source:** `plugins/core-sdlc/src/commands/review.ts`

### `lev loop`

Entity queue management — discover, validate, promote, budget.

```bash
lev loop [--json]              # Show entity queue
lev loop autodev [flags]       # Start heartbeat loop
```

**Exports from `@lev-os/orchestration`:** `discoverPlans`, `loadLoopConfig`, entity types.
**Source:** `plugins/core-sdlc/src/commands/loop.ts`

### `lev instruct`

Append structured instructions (rules, learnings, corrections) to a project.

```bash
lev instruct "rule text" [--source=human|agent]
```

Append-only to `.lev/instructions.md`. Never rewrites.
**Source:** `plugins/core-sdlc/src/commands/instruct.ts`

## Entity Lifecycle

The SDLC plugin manages entities through this state machine:

```
draft → ready → in_progress → needs_validation → validated → done
                    │              │
                    │              └─ (fail) → ready (with feedback)
                    └─ (blocked) → blocked
```

Entity types: `plan-impl`, `plan-chore`, `plan-bugfix`, `plan-research`, `spec`, `chore`

## Exec Profiles

| Profile | Purpose | Model |
|---------|---------|-------|
| `sdlc.flowmind.exec` | Execute + validate entity | opus-4-6 + codex |
| `sdlc.flowmind.deepen` | Deep plan decomposition | opus-4-6 |
| `sdlc.flowmind.hygiene` | Periodic health scan | sonnet-4-6 |

## Configuration

Lives in `plugins/core-sdlc/config.yaml` under the `sdlc` namespace:

```yaml
sdlc:
  paths:
    instructions: .lev/instructions.md
    evidence: .lev/evidence/
    specs: docs/specs/
  mayor_mode:
    enabled: false
  entity_definitions:
    plan: { ... }
```

Config cascade: system → project → module → env (later wins).

## Validation Gates

Spec validation includes `.lev/validation-gates.yaml` cross-reference. All SDLC checkpoints
(spec-lint, pre-execute, post-validate) check applicable enforced gates before proceeding.

## Capabilities

- `git_events` — Git hook event providers
- `bd_events` — Beads task tracking events
- `spec_linting` — Spec validation
- `instruction_capture` — Append-only instruction log
- `mayor_mode` — Notification suppression
- `exec_profiles` — SDLC-specific execution profiles

## Integration

- **Control plane hierarchy (D8):** `lev exec` is THE execution plane, `lev loop` is scheduling, `.flow.yaml` is policy
- **Skills are shims (D12):** `/work`, `/autodev-loop` are Claude Code skill shims — real functionality lives in `lev sdlc *` commands
- **Entity queue:** `lev loop --json` returns the priority-sorted entity queue

## Related Skills

- `/work` — FSM lifecycle router (Claude Code skill shim)
- `/autodev-loop` — Cron-based tick scheduler (Claude Code skill shim)
- `/autodev-lev` — Heartbeat-based tick scheduler via `lev loop autodev`
- `lev` — Master CLI reference

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…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.