Embrace the Agentic Development Lifecycle from inside Claude Code — phase-routing discovery skill plus commands to bootstrap and author ADLC tickets. Wraps the @adlc/cli gate toolkit.
Distill repeated review findings and PR rejections into permanent, deterministic defenses (ADLC P7).
Bootstrap the ADLC runtime (.adlc/) in this repo and verify the toolkit is installed.
Run the decay-driven ADLC maintenance checks — stale skills, hot files to re-prosecute, stale tickets, and gate calibration (C10/C12).
Prosecute the change for the active ticket. Prerequisite: a clean G4 build (rails
Author and triage an ADLC ticket (P0) through the canonical ticket store, then check it is executable.
P5 verifier subagent; invoked by /adlc:adlc-prosecute — do not invoke directly.
P5 contract lens subagent; invoked by /adlc:adlc-prosecute — do not invoke directly.
P5 correctness lens subagent; invoked by /adlc:adlc-prosecute — do not invoke directly.
P5 spec-vs-implementation lens subagent; invoked by /adlc:adlc-prosecute — do not invoke directly.
P5 security lens subagent; invoked by /adlc:adlc-prosecute — do not invoke directly.
Executes bash commands
Hook triggers when Bash tool is used
Modifies files
Hook triggers on file write and edit operations
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
Uses power tools
Uses Bash, Write, or Edit tools
Uses power tools
Uses Bash, Write, or Edit tools
A workspace of zero-dependency, gate-shaped Node.js CLIs for running software development the way frontier models actually behave — not the way humans do.
The SDLC is 60 years of defenses against human failure modes (forgetfulness, ego, fatigue). Models have a different flaw profile: premature satisfaction, sycophancy, context rot, confident hallucination, reward hacking. The Agentic Development Lifecycle (ADLC) redesigns every phase, gate, and loop around those flaws. Read the full thesis in ADLC.md.
Every phase, gate, and loop in this lifecycle must trace to a specific model failure mode it defends against, or a specific model property it exploits. If it traces to a human failure mode instead, cut it.
Each tool is a small CLI that enforces one machine-checkable gate. They share a runtime
convention (.adlc/ for tickets, ledgers, and gate evidence) and a common contract
(see CONVENTIONS.md) so 22 independently built tools feel like one
product.
Each package publishes independently under the @adlc npm scope. For normal use,
install the dispatcher and run tools through the stable adlc <tool> surface:
npm install -g @adlc/cli
adlc spec-lint <spec.md>
# or run without installing
npx @adlc/cli spec-lint <spec.md>
Individual packages can still be installed for package development, but Codex skills, hooks, CI, and lifecycle docs should use the dispatcher.
| Phase | Packages |
|---|---|
| Spec & ticket shaping | parallax · spec-lint · premortem · coldstart · ticket-prune |
| Execution supervision & rails | preflight · model-router · merge-forecast · rails-guard · build-gate · flail-detector · consensus-fix · runner |
| Review evidence & calibration | behavior-diff · gate-manifest · hollow-test · prosecute · review-calibration · model-ratchet · gate-fuzzing |
| Compounding defenses | lesson-foundry · rejection-mining · skill-rot |
| Shared foundation | @adlc/cli · @adlc/core |
See docs/package-reference.md for binaries, command forms, and per-package detail, and docs/toolkit.md for how the packages fit the ADLC flow.
This repo is also a Claude Code plugin — it makes the whole lifecycle usable from
inside the editor (phase-routing skill, ticket/distill/maintain commands, a
prosecutor subagent, and hooks that fire the gates automatically), with no API
keys (Claude is the model via --prompt-only).
Recommended: install with plugins, the
vendor-neutral installer that auto-detects your agent tools and installs the plugin
into each (Claude Code today; other harnesses as their plugin packages publish — each
has a native install path below; see ADR-0009
for why this is the preferred path):
npx plugins add voodootikigod/adlc # install the plugin into your agent tool(s)
npm install -g @adlc/cli # the gate toolkit the plugin shells out to
/adlc:adlc-init # bootstrap .adlc/ in your repo (once)
Already using Claude Code's native plugin marketplace? That path still works:
npm install -g @adlc/cli
/plugin marketplace add voodootikigod/adlc
/plugin install adlc@adlc
/adlc:adlc-init
See docs/integrations/claude-code.md for the full adoption guide.
Each agent tool has its own native integration. See the guides:
npx claudepluginhub voodootikigod/adlc --plugin adlcAccess thousands of AI prompts and skills directly in your AI coding assistant. Search prompts, discover skills, save your own, and improve prompts with AI.
Harness-native ECC plugin for engineering teams - 67 agents, 279 skills, 94 legacy command shims, reusable hooks, rules, MCP conventions, and operator workflows for Claude Code plus adjacent agent harnesses
Consult multiple AI coding agents (Gemini, OpenAI, Grok, Perplexity, plus codex, antigravity, and grok CLIs when installed) to get diverse perspectives on coding problems
Production-grade engineering skills for AI coding agents — covering the full software development lifecycle from spec to ship.
Evidence-gated AI coding workflow: scan → analyze → plan → TDD → execute → fix → verify → review, powered by Codebase Memory MCP >= 0.9.0 with optional Serena LSP intelligence. Includes blast-radius planning, test/cycle gates, independent review, and Windows Git Bash hook auto-resolution.
Feature development with code-architect/explorer/reviewer agents, CLAUDE.md audit and session learnings, and Agent Skills creation with eval benchmarking from Anthropic.