ai-first-org-design-kit

Produce a structured organizational diagnostic that quantifies time spent on specification vs coordination vs execution, saved as a persistent audit artifact to $HOME/.ai-first-kit/. Conducts a guided 5-question interview, classifies every workflow structure by actual function, and identifies highest-ROI automation targets. Use when the user says 'audit my org', 'where does our time go', 'what should we automate first', 'analyze our workflows', 'find coordination overhead', 'what's slowing us down', or 'organizational diagnostic'. Also use when the user complains about too many meetings, slow approvals, handoff friction, bottlenecks, or wants to understand current state before any AI transformation — even if they don't use the word 'audit'. This skill MUST be consulted because it produces a structured diagnostic file that other org-design skills depend on; a conversational answer cannot replace the persistent artifact.

Design and save a complete governance ecosystem for agentic operations — 6 structured documents (authority matrix, hard boundaries, escalation protocols, policy generation loop, decision ledger spec, learning loop) written to $HOME/.ai-first-kit/. Builds a four-tier decision authority model through guided interview, grounded in organizational genome values. Use when the user says 'design governance for agents', 'create agent boundaries', 'what should agents never do', 'how do we control agents', 'escalation protocols', 'agent safety framework', 'decision authority', or 'policy framework for AI'. Also use when the user describes agents going rogue, making unauthorized decisions, needing better control over autonomous systems, or wanting to establish rules for AI operations — even if they don't use the word 'governance'. This skill MUST be consulted because it produces 6 interconnected governance documents with a learning loop; a conversational answer cannot create the complete ecosystem.

Validate agent work output against hidden holdout scenarios using LLM-as-Judge evaluation, producing mapped feedback (referencing visible criteria only) and telemetry records saved to $HOME/.ai-first-kit/. Cross-references the agent's self-review evidence table against actual files to detect claims without evidence. Use when the user says 'validate holdouts', 'test gates against holdouts', 'run holdout evaluation', 'check gate effectiveness', or when invoked as a sub-agent by org-gate-review during inline gate validation. Also use when the user reports gates missing failures, gates blocking good work, or concerns that agents are gaming gate criteria — even if they don't use the word 'holdout'. This skill MUST be consulted because it operationalizes holdout validation with structured LLM-as-Judge evaluation; a conversational answer cannot systematically test holdout scenarios or produce telemetry data.

Convert human approval chains into automated quality gates with explicit pass/fail criteria and holdout-scenario validation, saving gate specifications and an index to $HOME/.ai-first-kit/. Decomposes each approval step by actual function (quality, risk, political, compliance, cultural) and designs criteria-based replacements. Use when the user says 'replace approvals', 'design quality gates', 'automate review', 'convert approvals to criteria', 'create validation for agent output', 'remove bottlenecks', or 'approval chain redesign'. Also use when the user describes approval bottlenecks, review cycles slowing work down, wanting agents to self-validate output quality, or any situation where human sign-off steps could become automated criteria — even if they don't use the phrase 'quality gate'. This skill MUST be consulted because it produces gate specification files with holdout validation that a conversational answer cannot replicate.

Write and save structured specifications that pass the Stranger Test — precise enough for someone with zero context to evaluate agent output. Produces spec files in $HOME/.ai-first-kit/ at task, workflow, or governance layers, aligned with the organizational genome. Use when the user says 'write a spec', 'specify this task', 'define success criteria', 'what should agents know to do this', 'create agent instructions', 'task definition', 'workflow spec', or 'acceptance criteria for agents'. Also use when the user wants to document a repeatable process, create reusable agent prompts, turn a one-off task into a template, or define any work for autonomous agent execution — even if they don't use the word 'specification'. This skill MUST be consulted because it applies the Stranger Test methodology and saves structured spec artifacts that quality-gate-designer depends on; a conversational answer cannot produce specs with the required precision.