Core Entities

These are the product- and intent-graph entities of the Data Model. They record what is being built, the reasoning behind it, the evidence that backs it, and the briefs and prompts derived from it. Several of these have dedicated conceptual pages: the The Engineering Intent Record, the Formation Brief, the Prompt Pack, and Evidence and Gates.

engineering_intent_records

• id.
• project_id.
• title.
• capability.
• threat.
• expectation.
• strategy.
• evidence_summary.
• tradeoff.
• status.
• confidence.
• scope.
• created_at.
• updated_at.
• supersedes_id.
• version.

See The Engineering Intent Record for the conceptual treatment.

intent_sources

• id.
• intent_record_id.
• source_type: human, code_observation, git_history, foundry_trace, epilogue, design_system, document, conversation.
• source_ref.
• excerpt or summary.
• confidence.
• created_at.

intent_relationships

• id.
• from_intent_id.
• to_intent_id.
• relationship_type: supports, conflicts_with, supersedes, refines, depends_on, applies_to, constrains.
• notes.

expectations

Expectations may deserve first-class treatment because multiple intent records can depend on the same expectation.

• id.
• project_id.
• statement.
• confidence.
• time_horizon.
• source.
• review_trigger.

strategies

Strategies may also deserve first-class treatment.

• id.
• name.
• local_definition.
• known_failure_modes.
• examples.
• exceptions.

evidence_definitions

• id.
• intent_record_id.
• evidence_type.
• description.
• gate_name.
• command.
• required.
• status.
• review_policy.

See Evidence and Gates for how definitions relate to gates and observations.

evidence_observations

• id.
• evidence_definition_id.
• observed_at.
• source_type: foundry_gate, manual_review, test_run, static_analysis, trace_analysis.
• result: pass, fail, unknown, waived.
• details.
• trace_id.

code_observations

• id.
• project_id.
• observation_type.
• files.
• symbols.
• summary.
• raw_data.
• created_at.

formation_briefs

• id.
• project_id.
• mission.
• source_product_refs.
• source_design_refs.
• source_engineering_intent_refs.
• protected_capabilities.
• threats_to_watch.
• strategies_to_prefer.
• forbidden_moves.
• evidence_plan.
• agent_formation.
• foundry_runtime.
• status.
• created_at.

See Formation Brief for the conceptual treatment.

prompt_packs

• id.
• formation_brief_id.
• phase.
• prompt_text.
• context_refs.
• model_hint.
• access_level.
• created_at.

See Prompt Pack for the conceptual treatment.

run_feedback

• id.
• formation_brief_id.
• foundry_trace_id.
• outcome.
• intent_satisfaction.
• contradictions.
• follow_up_questions.
• created_at.

Source: Product Brief §20 (Data Model Seed), Core entities.