super-build-kit

name	super-build-kit
description	Use when the task involves any build, fix, audit, debug, optimization, release, verification, prompt, skill, or agent-design task. Activates the APIVR lifecycle and Elite Build Goals guardrails from this repository before task-specific work.
activation	Activate when the description trigger applies to the current task.
required_inputs	Task request, relevant repository context, constraints, and authority dependencies.
required_outputs	Skill-specific artifact, verification evidence, canonical verdict, and next action.
authority_dependencies	00_start_here/SOURCE_OF_TRUTH.md; 10_governance/APIVR_EXECUTION_LIFECYCLE.md; 10_governance/source_of_truth/Elite_Build_Goals_v3.md.
evidence_requirements	Executed checks or an honest Unknown, Not Run, or Blocked state for every material claim.

Super Build Kit

This skill is the first orientation layer for the portable Super Build Kit operating system. Use it to decide which APIVR files, skills, templates, and specialist roles must govern the task.

Required Activation

Read these files in order:

1. 00_start_here/START_HERE.md
2. 00_start_here/SOURCE_OF_TRUTH.md
3. 00_start_here/LOAD_ORDER.md
4. 50_audits/AUDIT_TIER_ROUTER.md

Then load only the task-specific files named by LOAD_ORDER.md.

Skill Invocation Flow

flowchart TD
  A["User asks for work"] --> B["Load source of truth and APIVR tier router"]
  B --> C{"Is this planning or handoff?"}
  C -- "Yes" --> D["Load writing-plans"]
  C -- "No" --> E{"Will code change?"}
  E -- "Yes" --> F["Load test-driven-development"]
  E -- "No" --> G["Use evidence-first verification"]
  D --> H{"Can work be safely isolated or delegated?"}
  F --> H
  G --> H
  H -- "Worktree needed" --> I["Load using-git-worktrees"]
  H -- "Subagents useful" --> J["Load dispatching-parallel-agents and subagent-driven-development"]
  H -- "No" --> K["Proceed sequentially"]
  I --> L{"Is work long-running, tool-heavy, recurring or iterative?"}
  J --> L
  K --> L
  L -- "Long-running/tool-heavy" --> M["Load long-horizon runtime and run tracing"]
  L -- "Recurring/iterative" --> P["Load repeatable-agent-loops"]
  L -- "No" --> Q{"Security-sensitive or dual-use?"}
  M --> N
  P --> N
  Q -- "Yes" --> R["Load cybersecurity risk routing and security skills"]
  Q -- "No" --> N["Load domain skills by task"]
  R --> N
  N --> O["Execute APIVR and report verdict"]

For implementation plans, feature work, refactors, fixes, or risky edits:

• Load skills/writing-plans/SKILL.md before plan creation.
• Load skills/20-pass-protocol/SKILL.md before finalizing high-stakes prompts, agents, skills, source files, plans, audits, runbooks, templates, release instructions, or other accuracy-critical artifacts.
• Load skills/test-driven-development/SKILL.md before APIVR Phase 3 code work.
• Load skills/using-git-worktrees/SKILL.md and prefer native worktree tooling before manual git fallback.
• Load skills/dispatching-parallel-agents/SKILL.md and skills/subagent-driven-development/SKILL.md before dispatching delegated work. skills/subagent-driven-development/SKILL.md owns the deterministic controller protocol: pre-flight conflict scan, durable task briefs, exact base..head review packages, mandatory fix and re-review of material findings, bounded repair budgets, and a final independent whole-branch review.