7. Ceremonies

Ceremonies handle human strategic decisions. Quality enforcement is handled by Gates (Section 8).

7.1 Increment Planning (SHALL)

Define objectives, task pool, budget, and risks for the upcoming Increment.

Outputs: objectives, prioritized tasks, planned budget, risk register.

7.2 Session Start (SHALL)

Load context, select tasks, verify environment.

7.3 Session End (SHALL)

Capture metrics, write handoff, record knowledge.

SHALL outputs: summary, task states, handoff (next steps, warnings), metrics, dead ends (if any). SHOULD outputs: knowledge entries for decisions/patterns/known issues.

7.4 Quality Sweep (SHALL)

Periodic comprehensive audit of codebase and knowledge base.

Scope: security, code quality, test health, configuration drift, knowledge freshness, architectural conformance, dependency health, AI-specific issues (duplication, scope creep accumulation, TODO debris).

7.5 Federation Sync (SHOULD; SHALL at Team+)

Coordinate multiple Supervisor+AI Pairs.

Agenda: status per Pair, dependency updates, shared component changes, integration issues.

7.6 Delivery Review (SHOULD)

Demonstrate working software to stakeholders, collect feedback, obtain acceptance.

7.7 Increment Retrospective (SHALL)

Quantitative review: planned vs actual cost, throughput trend, FPSR, DER, ADR, knowledge capture rate.

Output: specific, measurable, time-bounded, assigned improvement actions.

Organizations SHOULD allocate time for innovation, learning, and AI tooling experimentation.

7.8 Summary