Skip to main content

Glossary

Use this page when VibeGov language starts to feel dense.

Definitions here are short on purpose. Follow the linked docs when you need the fuller rule, workflow, or bootstrap contract.

Core delivery terms​

Governance layer​

A repo-local set of rules, project intent, specs, workflow expectations, and evidence standards that guide how humans and agents deliver software. See Overview and The VibeGov SDLC.

Source of truth​

The canonical place a project expects agents and humans to read from before acting. In VibeGov, .governance/ is the governance source of truth. See Overview.

Canonical contract​

The single contract a workflow must satisfy rather than a loose set of optional suggestions. Bootstrap uses one canonical contract across init, update, and review modes. See Bootstrap.

Evidence​

The proof that a claim is true: commands run, checks passed, artifacts created, screenshots captured, issues updated, or blockers recorded. See Checkpoint Reporting and Test Execution Expectations.

Traceability​

The connection between intent, issue, spec, implementation, validation, and follow-up work. Traceability keeps future changes understandable. See The VibeGov SDLC.

Spec binding​

The act of linking work to a concrete spec, requirement, or explicit SPEC_GAP before claiming delivery confidence. See Mode Selection and Evidence Closing.

Modes and work shapes​

Development​

The mode for changing reality: building, fixing, validating, preparing release evidence, and leaving a reviewable delivery trail. See Execution Modes.

Exploration​

The mode for discovering reality: review, audit, judgment, gap finding, and backlog hydration without pretending a fix has been delivered. See Execution Modes and Exploratory Review Mode.

Feedback Intake​

The governed work shape for turning human feedback into ready issues, specs, priority notes, or dependency notes before implementation starts. See Feedback Intake.

Evaluation​

A criteria-based judgment pattern used inside Development or Exploration. It is not a third peer mode. See Evaluation Pattern.

Backlog hydration​

Turning findings, feedback, defects, or missing requirements into focused backlog items with enough context to execute later. See Exploratory Review Mode.

Bootstrap terms​

Bootstrap​

The process of installing or checking the repo-local governance structure before product-code implementation. See Bootstrap.

Bootstrap init​

Bootstrap mode for creating missing governance state in a repo. See Quick Start.

Bootstrap update​

Bootstrap mode for repairing or normalizing existing governance state against the same contract. See Bootstrap Update.

Bootstrap review​

Bootstrap mode for auditing a repo against the same contract without claiming missing work was completed. See Bootstrap Review.

Pass gate​

The minimum set of conditions that must be true before a workflow can honestly claim it passed. Bootstrap has a documented pass gate. See Bootstrap.

Role pack​

A reusable bundle of operating instructions for an agent or role in a governed project. See Agents Template.

Harness profile​

Guidance for mapping a specific agent harness or runtime pattern onto VibeGov controls without making that tool the governance source of truth. See Harness Profile: Codex and Harness Profile: Minimal Claude Harness.

Evidence and completion terms​

Checkpoint​

A visible progress update or artifact that records the current state, evidence, risk, blocker, or next action while work is still recoverable. See Checkpoint Reporting.

Completion claim​

A statement that work is done. In VibeGov, completion claims need mode-appropriate evidence, not just confidence or output. See Mode Selection and Evidence Closing.

Release verification​

Build, startup, smoke, packaging, deployment, or release checks used to prove a candidate is safe enough to ship. Release verification belongs inside Development. See Execution Modes.

Residual risk​

Known uncertainty left after validation. It should be named instead of hidden behind a successful summary. See Workflow Quality Rubric.

Agent and continuity terms​

Continuity​

The durable state that lets future work resume without depending only on live chat memory. See Agent Continuity Bootstrap.

Blocker​

A proven condition that prevents meaningful progress or confidence in a scoped item. A blocker should be recorded, routed, and bounded. See Blocker Escalation.

Scoped blocking​

The practice of pausing only the blocked item or surface while keeping unrelated ready work moving when possible. See Build Loop, Exploratory Loop, Human Feedback Loop, and Scoped Blocking.

Agent legibility​

The quality of making agent work understandable through clear state, durable evidence, and recoverable handoffs. See Published GOV 11.

Drift​

The gap that opens when implementation, intent, evidence, specs, or issue state stop matching each other. See Published GOV 12.

Quick map​