Agent skills that bring team coding standards to Claude Code and Codex

Stop Vibe Coding successful Silos. Build a Shared Cognitive Layer for Your Engineering Team.

Individual punctual hacks create speedy wins for solo developers, but erstwhile scaled crossed a team, "vibe coding" leads to chaotic method debt, discourse rot, unreviewable PRs, and mislaid codification ownership. Speed is solved; Trust and Verification is the caller bottleneck successful AI engineering.

ADLC Team Skills (tikalk/adlc-team-skills) is the open-source Team Layer of the Twelve-Factor Agentic SDLC. It turns AI agents from isolated guessers into compliant, accountable squad members that stock your team's constitution, merchandise strategy, architectural standards, and information benchmarks.

Team AI Directives (tikalk/agentic-sdlc-team-ai-directives) is the companion repository that holds your team's version-controlled discourse modules (constitution, rules, personas, examples), CDR index, and skills manifest.

Works retired of the container pinch immoderate supplier supporting the Agent Skills standard — Claude Code, Codex, OpenCode, Cursor, GitHub Copilot, and others.

Slash commands + events: adlc-skills-cli wraps npx skills adhd and additionally generates /name slash commands and wires session_start arena hooks (via .events.json) for 9 coding agents. Skills repos without .events.json get commands only.

Universal orchestration: mission-brief auto-discovers skills from any root (mattpocock/skills, addy osmani/agent-skills, superpowers, spec-kit, aliases your own) and dynamically wires them into the ngo pipeline. No vendor lock-in.

team-boot auto-runs astatine convention commencement via the arena hook. On an unconfigured task it outputs a informing telling the personification to tally /team-setup. team-setup is besides disposable connected demand:

Then take Mode 3 — Scaffold caller quiet team-ai-directives, aliases Mode 1 — Clone from GitHub to fork tikalk/agentic-sdlc-team-ai-directives:

Already person a directives repo? team-setup offers 3 different modes:

• Mode 1 — Clone from GitHub (e.g. fork tikalk/agentic-sdlc-team-ai-directives)
• Mode 2 — Point to existing section path (wire a repo you already have)
• Mode 4 — Already configured (verify an existing setup)

Factor I — Developer arsenic Orchestrator. Factor X — Context Engineering. Factor XI — Directives arsenic Code.

Put the team astatine the halfway of your AI strategy. Instead of individual developers hoarding punctual shortcuts connected section machines, squad standards unrecorded successful a version-controlled Git repository (team-ai-directives).

• team-boot: Auto-runs astatine convention commencement via the arena hook, assembling the squad constitution, CDR index, PDR/ADR indexes, and accomplishment registry into the strategy prompt.
• team-discover: Manually re-scan squad discourse modules for definitive system find tables. Available via /team-discover.
• team-constitution: Interactively define, review, aliases amend your engineering team's halfway principles.
• team-repair: Re-index CDR.md, scan for norm conflicts, and verify directive freshness.

Factor III — Mission Definition. Factor IV — Structured Planning. Factor IX — Traceability.

Without documented decisions, each implementation convention re-derives (or misinterprets) merchandise intent and architectural rules.

• Product Decision Records (product-*): Capture merchandise decisions arsenic individual PDR files, resoluteness ambiguities done an interactive explanation workflow, and compile them into a self-contained PRD.md.
• Architectural Decision Records (architect-*): Reverse-engineer aliases specify architectural decisions utilizing Rozanski & Woods viewpoints (Functional, Security, Deployment, Performance) and constitute them into a unified AD.md.
• product-roadmap: Track milestone advancement crossed 4 layers of truth — decisions (PDRs), execution (live issues via MCP), codification evidence, and milestone gates.

Factor III — Mission Definition. Factor IV — Structured Planning. Factor V — Triage & Execution. Factor XIII — Loop Engineering.

AI is an obsessive guesser — erstwhile faced pinch ambiguity, it invents solutions alternatively of asking questions. Move from a Conversational exemplary to a Contract model.

• mission-brief: The team's autonomous pipeline runner. Takes a characteristic prompt, derives a general statement (Goal, Constraints, Non-Goals, Success Criteria), generates an ordered measurement list, and walks a specify → scheme → tasks → instrumentality ↺ converge loop to completion.
• Mantra: "Debug the Spec, Not the Code": When an supplier makes a mistake, don't conscionable spot the codification — adhd the missing constraint to the specification truthful the correction is ne'er repeated.

Factor VII — Verification-First Evals. Factor VIII — Ratchet Effect. Factor IX — Traceability. Factor XII — Build to Delete.

Never fto the supplier that wrote the codification determine if the codification is good. "Separate the Maker from the Checker."

• evals skills: Build application-level information suites (PromptFoo aliases DeepEval) utilizing Eval-Driven Development (EDD). Runs Tier 1 accelerated checks + Tier 2 LLM judge subagents to trial codification against defined business risks before quality macro-review successful The Great Filter.
• levelup: Capture convention wins into imperishable squad memory. levelup-specify extracts convention execution traces and commits them to Git arsenic reusable rules (Context Directive Records — CDRs).
• "Build to Delete": Prune outdated rules and punctual scaffolding arsenic underlying instauration models amended utilizing team-repair --build-to-delete.

Factor XII — Build to Delete. Factor XIII — Loop Engineering.

mission-brief acts arsenic an open, vendor-agnostic orchestrator crossed each celebrated supplier accomplishment repositories and Spec-Driven Development (SDD) frameworks:

1. Discovery — At ngo start, mission-brief scans each skills directories (.claude/skills, .agents/skills, etc.) and sounds each SKILL.md frontmatter to build a vendor-agnostic inventory of installed skills pinch their names and descriptions.
2. LLM-decided routing — Each step's delegation punctual includes the afloat skills inventory. The subagent decides which accomplishment (if any) fits the existent shape — the LLM matches, not a brittle lookup table.
3. Graceful fallback — If nary accomplishment matches, the subagent executes directly. If a accomplishment matches, it's invoked. Either way, the ngo pipeline continues.

Skills are organized nether the 4 pillars of the Twelve-Factor Agentic SDLC, flattened straight nether the skills/ directory:

This places each azygous accomplishment precisely 2 levels deep, afloat resolving the default extent limit of the skills CLI and ensuring each skills instal retired of the box.

• team-boot — Bootstrap session: assembles constitution, CDR index, PDR/ADR indexes, and accomplishment registry into the strategy punctual astatine convention start. Outputs a informing to tally /team-setup connected unconfigured projects.
• team-discover — Manually re-scan squad discourse modules and nutrient a system find table. Available via /team-discover.
• team-setup — Clone, scaffold, aliases configure a squad AI directives repository. Model-invoked by team-boot (self-install) and disposable connected demand. Say "Set up squad directives for this project."

• team-constitution — Create aliases amend the squad constitution interactively. Say "Create our squad constitution" aliases "Amend our squad principles."
• team-repair — Re-index CDR.md, .skills.json, AGENTS.md; wellness check; conflict scan; freshness verification. Say "Check our squad directives health" (--health-only), "Repair our CDR index," aliases "Scan for norm conflicts" (--conflicts).
• team-skills — Browse and instal squad skills from the squad AI directives. Say "Show maine disposable squad skills."

All user-invoked. Capture and people reusable patterns to team-ai-directives, including paired directive compliance evals.

• levelup-init — Brownfield CDR find from existing codebase, including paired eval CDRs from codification patterns. Say "Discover directives from this codebase."
• levelup-specify — Extract CDRs and paired eval CDRs from the existent session. Say "Extract lessons from this session."
• levelup-clarify — Review, accept, reject, aliases defer pending CDRs. Evals regression gross runs by default. Say "Review pending CDRs."
• levelup-publish — Compile accepted CDRs into squad directives artifacts, evals goldensets, and draught PR. Say "Publish accepted CDRs" aliases "Build 1 accomplishment from a CDR" (--skill CDR-NNN).

All user-invoked. Document merchandise decisions arsenic individual PDRs and compile into a self-contained PRD.md.

• product-init — Brownfield PDR find from existing codebase and documentation. Say "Discover merchandise decisions from this codebase."
• product-specify — Greenfield PDR creation done interactive merchandise exploration. Say "Let's specify our merchandise strategy."
• product-clarify — Refine, validate, and o.k. PDRs earlier PRD generation. Say "Review our merchandise decisions."
• product-implement — Generate PRD.md from accepted PDRs (multi-agent DAG orchestration). Say "Generate our PRD."
• product-analyze — Read-only PDR↔PRD consistency and value analysis. Say "Analyze our merchandise docs."
• product-roadmap — Track milestone progress: determination status, unrecorded issues via MCP, codification evidence, and gates. Say "Show roadmap progress."

All user-invoked. Create and negociate Architecture Decision Records utilizing the Rozanski & Woods methodology.

• architect-init — Reverse-engineer ADRs from an existing codebase (brownfield). Say "Reverse-engineer architecture from this codebase."
• architect-specify — Create ADRs from a PRD aliases characteristic explanation (greenfield). Say "Create ADRs from this PRD."
• architect-clarify — Refine and validate existing ADRs. Say "Refine and validate my ADRs."
• architect-implement — Generate an Architecture Description (AD.md) from accepted ADRs. Say "Generate AD.md from my ADRs."
• architect-analyze — Check ADR↔AD consistency and architecture quality. Say "Analyze architecture consistency."

All user-invoked. Build and support application-level information suites pursuing EDD (Eval-Driven Development) principles (PromptFoo aliases DeepEval).

• evals-init — Initialize information directory building (evals/{system}/) pinch information baseline. Say "Initialize my information harness."
• evals-specify — Extract eval criteria from specs and accumulation nonaccomplishment traces (bottom-up unfastened coding). Say "Specify information criteria from this nonaccomplishment log."
• evals-clarify — Cluster related patterns, refine criteria, isolate 20% holdout split, and people goldset. Say "Clarify and judge my draught evaluations."
• evals-implement — Generate executable graders and trial configs, automatically moving portion tests to verify evaluator correctness. Say "Generate graders from the goldset."
• evals-validate — Run the information pyramid (Tier 1 accelerated checks + Tier 2 LLM judges) and compute value metrics (TPR/TNR, SLA headroom). Say "Validate my information suite."
• evals-analyze — Deep-analyze trajectory nonaccomplishment traces, routing spec-level failures to levelup-specify (rules) and generalization failures to backlog. Say "Analyze information failures."

User-invoked. Structure a characteristic explanation into a Mission Brief and tally it end-to-end pinch immoderate installed SDD accomplishment set.

• mission-brief — Takes a description, structures it into a Mission Brief (goal, constraints, occurrence criteria), generates an ordered measurement database pinch prompts that trigger installed SDD skills, and walks those steps to converged implementation. Sync (gated) aliases --async (ungated, checkpoint crossed sessions). Say "Build this characteristic extremity to end" aliases mission-brief "add acheronian mode". Resume pinch mission-brief --resume.

Model-invoked. Grounds tech stack choices successful Tikal's Israeli Tech Radar.

• tech-radar-context — Discovers technologies implied by the prompt, matches them against the Tikal Tech Radar (radar.json), and injects a discourse array pinch each technology's take ringing (Keep/Start/Try/Stop), quadrant, and Tikal's "Why?" sentiment — positive Tikal-aligned alternatives for thing connected Stop. Auto-triggered whenever a technology, framework, database, library, aliases unreality instrumentality is being chosen aliases evaluated. Fetches the unrecorded radar best-effort and falls backmost to a bundled snapshot astatine resources/radar.json.

All skills constitute to .adlc/ (project root) and the squad AI directives repo.

Team Directives (inside the squad AI directives repository):

• AGENTS.md — supplier instructions (loading order, rules, skills)
• CDR.md — scale of approved discourse contributions
• .skills.json — skills manifest (schema v2.0.0)
• .mcp.json.example — MCP servers config example
• context_modules/constitution.md — squad constitution (OKF frontmatter)
• context_modules/{rules,personas,examples}/**/*.md — discourse modules
• context_modules/{type}/index.md — progressive disclosure per conception type
• context_modules/{type}/log.md — chronological alteration log per conception type
• skills/{name}/SKILL.md + .skills-entry.json — published squad skills
• evals/{directive-id}/goldset.md + goldset.json — directive compliance goldensets

LevelUp (inside .adlc/ of the target project):

• .adlc/drafts/cdr/CDR-{NNN}.md — proposed/discovered CDRs (including eval CDRs)
• .adlc/drafts/cdr/cdr.md — auto-generated CDR index
• .adlc/init-options.json — squad AI directives way config

Product (inside .adlc/ and repo root):

• .adlc/drafts/pdr/PDR-{NNN}.md — proposed/discovered PDRs
• .adlc/drafts/pdr/pdr.md — auto-generated PDR index
• .adlc/memory/pdr/PDR-{NNN}.md — accepted/completed PDRs
• .adlc/memory/pdr/pdr.md — accepted PDR index
• .adlc/product/sections/{feature-area}/{section}.md — PRD conception build artifacts
• .adlc/product/state.json — DAG execution state
• PRD.md — Product Requirements Document (repo root)

Architecture (inside .adlc/ and repo root):

• .adlc/drafts/adr/ADR-{NNN}.md — proposed/discovered ADRs
• .adlc/drafts/adr/adr.md — auto-generated ADR index
• .adlc/memory/adr/ADR-{NNN}.md — accepted ADRs
• .adlc/memory/adr/adr.md — accepted ADR index
• AD.md — Architecture Description (repo root)
• .adlc/architect/ — per-view DAG artifacts

Missions (inside .adlc/ of the target project):

• .adlc/workflow/workflow-config.yml — ngo execution/supervision/budgets config
• .adlc/workflow/.mission-state.json — measurement list, completed steps, brief, find results
• .adlc/workflow/runs/<feature>/mission-log.json — last audit trail
• .adlc/workflow/runs/<feature>/iterations.md — per-implement audit entries

Governance (inside target task and repo root):

• .adlc/drafts/evals/EVAL-{NNN}.md — proposed/discovered eval criteria drafts
• .adlc/drafts/evals/evals.md — draught evals index
• .adlc/memory/evals/EVAL-{NNN}.md — accepted/completed eval criteria
• .adlc/memory/evals/evals.md — accepted evals index
• .adlc/memory/evals/holdout.json — isolated/reserved holdout trial dataset
• evals/{system}/goldset.md — published goldset (human-readable)
• evals/{system}/goldset.json — published goldset (machine-readable)
• evals/{system}/config.yml — information model configuration
• evals/{system}/config.{js,py} — model trial config
• evals/{system}/graders/check_*.py — generated binary Python graders / metrics
• evals/{system}/tests/test_check_*.py — generated portion tests verifying grader correctness
• evals/results/validation_report.md — statistical validation results report

Generated discourse modules see Open Knowledge Format (OKF) v0.1 compliant frontmatter alongside civilization fields.

Custom fields co-exist pinch OKF frontmatter: id, cdr_ref, created, modified, verified, age_days, evidence.

Directory structure: context_modules/{type}/index.md (progressive disclosure), context_modules/{type}/log.md (change history), cross-links betwixt related concepts.

Team Directives setup:

Product lifecycle:

Architecture lifecycle:

LevelUp / CDR lifecycle:

Mission:

Application Evaluation lifecycle:

Full merchandise → architecture → team:

See RELEASE.md for the merchandise runbook, tag naming conventions, and betterment procedures.

MIT — spot LICENSE.