The documents
13 numbered documents and 9 templates. Each states what it requires, what it recommends and what it leaves open, and each carries a stable identifier.
Further readingThe original Agentic Sprint essay on dalugoda.comAutonomous Loop, the optional inner-loop implementation
Agentic Sprint methodology
The normative core states what a conforming Agentic Sprint requires. The informative documents explain, measure and apply that core without adding requirements of their own. The working templates are the artefacts a team fills in while running a sprint. Autonomous Loop is an optional execution profile: a team can conform to Agentic Sprint without it.
Normative core
Requirements a conforming implementation must meet.
- D1. Agentic Sprint Specification v0.1A formal specification for human-governed software delivery with bounded autonomous execution, independent verification and explicit release authority.
- D4. Autonomous Loop Integration SpecificationA specification for using Autonomous Loop as the persistent, resumable execution protocol inside a human-gated Agentic Sprint.Builds on D1, D2, D3
- D5. Build Plan StandardA reusable standard for producing, reviewing and approving the plan required before agentic implementation begins.Builds on D1, D2, D3, D4
- D6. Agentic Definition of Ready and Definition of DoneA practical conformance model for deciding when an agentic work item may start and what evidence is required before implementation can be considered complete.Builds on D1, D3, D5
- D7. Agentic Engineering Guardrails and Security ModelA security model for bounded agent execution, independent verification, context integrity and human-controlled release in an Agentic Sprint.Builds on D1, D2, D4, D6
- D8. Agent Learning and Context Governance StandardA governance standard for turning human review corrections into durable, tested and reversible agent context.Builds on D1, D3, D4, D7
- D9. Agentic Sprint Quality and Verification ModelA layered verification model for challenging, accepting and releasing agent-produced software without confusing green checks with proof of correctness.Builds on D1, D4, D5, D6, D7, D8
- D10. Agentic Sprint Metrics FrameworkA balanced measurement framework for agentic delivery that connects flow, quality, economics, learning and safety without reducing engineering to velocity.Builds on D1, D3, D6, D8, D9
Informative documents
Guidance, measurement and argument that support the core without adding requirements.
- D2. Agentic Sprint Reference ArchitectureA reference architecture for moving approved software work through orchestration, bounded agent execution, independent verification, human review and controlled release.Builds on D1
- D3. Agentic Sprint Operating ModelHow a software delivery team can organise roles, decisions, ceremonies, work in progress and escalation around human-gated agent execution.Builds on D1, D2
- D11. Agentic Sprint ManifestoTen principles for human-governed software delivery with autonomous agents.Builds on D1, D7, D8, D9, D10
- D12. Agentic Sprint Implementation GuideA practical 30, 60 and 90 day path from AI-assisted engineering to governed agentic delivery.Builds on D1, D2, D3, D4, D5, D6, D7, D8, D9, D10, D11
- D13. Agentic Sprint Whitepaper v1.0A proposed operating model for human-governed software delivery with autonomous execution, independent verification and explicit release authority.Builds on D1, D2, D3, D4, D5, D6, D7, D8, D9, D10, D11, D12
Working templates
Artefacts to copy into a repository and fill in during a sprint. Released under CC0, so no attribution is required.
- T1. Build Plan TemplateA reusable Human Gate 1 plan for translating an approved requirement into bounded, reviewable engineering work.Builds on D1, D6
- T2. Agent Skill TemplateA reusable, reviewable definition for an agent skill with explicit scope, permissions, evidence and evaluation.Builds on D1, D8
- T3. Architecture Context TemplateA maintainable context record for giving agents a current view of system boundaries, invariants, decisions and ownership.Builds on D1, D2
- T4. Agent Handover TemplateA resumable handover record for preserving agent execution state, evidence, failures and the next safe action.Builds on D1, D4
- T5. Review Evidence TemplateA structured independent Checker record for testing implementation claims before Human Gate 2.Builds on D1, D9
- T6. Security Assessment TemplateA reusable security and authority assessment for agentic work, including assets, trust boundaries, tools, threats, controls and acceptance.Builds on D1, D7
- T7. Jira Story TemplateA reusable work-item record that gives an agentic delivery system clear intent, acceptance conditions, constraints and readiness evidence.Builds on D1, D6
- T8. QA Evidence TemplateA reusable independent product-assurance record for validating behaviour, journeys, regression risk and unresolved quality concerns.Builds on D1, D9
- T9. Agentic Sprint Retrospective TemplateA reusable retrospective record for examining flow, human intervention, quality, guardrails and durable learning after agentic delivery work.Builds on D1, D8, D10