ReferenceContracts

Reference

Command, marker, manifest, payload, and validation contracts for operators and maintainers.

Boundary

Runtime payload stays lean

Only execution-critical files install into a working project. Source-only validators, imported provenance, and site generation remain in the repository.

**Runtime payload diagram**Installed files are separated from repository validation and documentation sources.

Markers

Evidence marker contract

@EVIDENCE:{id}Observed source or input.

?UNKNOWN:{id}Material uncertainty.

#FINDING:{id}Evidence-backed conclusion.

+ACTION:{id}Bounded action or recommendation.

=VERIFY:{id}Acceptance check and result.

!STOP:{reason}Required halt condition.

Catalog

Generated library references

These pages are generated from canonical repository data so prompt, scenario, and pair explanations stay aligned with validation.

Prompts

Prompt catalog

Every canonical prompt with modes, contracts, tags, artifacts, and usage boundaries.

Scenarios

Scenario catalog

Atomic routes and composite workflows with graph, phase, lock, and gate detail.

Pairs

Prompt pairs

Planning/execution relationships and the handoff contract between reciprocal twins.

REFERENCECapability ArchitectureThe suite uses control prompts, standalone investigations, standalone production or operations, independent gates, and genuine investigate→execute or brief→produce pairs. Pairing i→ REFERENCECoverage IndexThe suite contains **257 prompts**, **32 genuine pairs**, **110 composite scenarios**, **27 department packs**, and **277 registered skills**. The supplied OpenCode snapshot contai→ REFERENCEEvidence LanesEvery task uses one lane. The lane controls how facts, interpretation, and invention may appear.→ REFERENCEGetting StartedMission Directives installs a command-line runtime globally and pins a complete operational copy inside each project. The project copy keeps routing behavior stable across global p→ REFERENCEPrompt Catalog**Prompt count:** **257**→ REFERENCEPrompt Engineering MethodsThe suite uses task-fit methods rather than adding every technique to every prompt.→ REFERENCEStructured Prompt Design StandardThis document defines the canonical body contract for every prompt in the suite. It is normative: prompt authors, transformation scripts, validators, scenario compilers, skill adap→ REFERENCEPrompt Suite ConventionsThis file defines suite-wide conventions that apply to every prompt, scenario, skill adapter, model adapter, runtime artifact, manual, and validator.→ REFERENCEDocumentation LibraryThe documentation library explains how to operate, author, validate, secure, and extend the prompt suite. It is organized by reader task rather than file age or generation order.→ REFERENCERecommended Skill StacksThese are the highest-value default stacks for this suite. Install only the exact skills needed for the current workflow, review the current source and audits, and keep native prom→ REFERENCEResearch BasisThe suite applies prompt-engineering techniques selectively according to task, model, risk, and output medium. No technique is treated as a universal incantation. The operating sta→ REFERENCERuntime Marker ProtocolRuntime markers give human-readable artifacts stable machine identifiers. They connect evidence, uncertainty, findings, actions, verification, and stops across prompts, tools, mode→