the receipts tradition · warmth you can verify
Research Library
Essays, Field Journal entries, and fieldcraft methods from the Hearth & Code studio — every claim labeled, every artifact traceable. Reading before commitment is always free.
Hearthside essays & Field Journal
Hearthside essay · N-05Meaning Before MachineryBefore a system is allowed to become clever, it should be able to say what it is for, whose situation it touches, and what it cannot decide.Hearthside essay · N-06The Craft of Keeping Uncertainty VisibleUncertainty does not have to be buried beneath confidence or inflated into paralysis. It can be given a clear place in the working surface.Hearthside essay · N-07A Humane Theory of ProvenanceProvenance is not paperwork added after the interesting work. It is the companion that helps a claim remain accountable when it leaves the room where it was made.Hearthside essay · N-08Tools in a Bounded RoomA useful agent is not an oracle or a colleague with invisible authority. It is a tool working inside a room whose doors, materials, and return path are clear.Technical article · N-09Typed Context for Real WorkTyped context is a small discipline for stopping sources, claims, proposals, decisions, and receipts from borrowing one another’s authority.Technical article · N-10Knowledge Representation Without a Black BoxA knowledge representation can be technical and still remain readable when it shows its sources, transformations, omissions, and human review points.Technical article · N-11Operational Intelligence as a Comparison PracticeOperational intelligence is useful when it compares forms, transitions, and boundaries without pretending to become the authority over every system it can describe.Technical article · N-12Prompting as Interface DesignA prompt is an interface for arranging attention: it selects materials, exposes boundaries, shapes a return, and leaves consequential choices with people.Field Journal · FJ-01A Hub That Can Explain ItselfAn eight-stage, illustrative tutorial for prototyping provenance-aware AI-assisted knowledge work, with explicit boundaries between candidate artifacts and a working system.Field Journal · FJ-02A Public Journal Begins With a BoundaryA note about keeping a public record honest: sharing methods, questions, and corrections without turning private work or one person's experiment into a general claim.Field Journal · FJ-03An Extended Mind Needs a GrammarA first-person N=1 field note on returnable AI-assisted work, durable records, bounded context, provenance, and human authority.Field Journal · FJ-04Designing a Genre as a SystemAn N=1 field note on designing a music genre as a system: the Exocore cosmic-ritual concept, four movements, CORE+TWEAK budgets, metatag grammar, lyric craft, and the contracts that make the genre.Field Journal · FJ-05ESS: The Thread Between Thought and ActionA Hearthside orientation to ESS: a candidate framework for carrying provenance, meaning, and human authority across AI-mediated work.Field Journal · FJ-06Every Claim Needs a Trail Back HomeA dated case study of the Governance Lab 0.0.26 Obsidian prototype, its source and effect boundaries, and what that historical account does not establish about current runtime status.Field Journal · FJ-07HCAN: A Condensed Symbolic Language for Human-Agent CommunicationA field-journal projection of HCAN: condensed symbolic language for human-agent communication, with worked examples, protocol benchmarks, evaluation context, and the portable envelope.Field Journal · FJ-08Prompting Became Procedure DesignAn N=1 field note on how prompting moved from clever phrasing toward procedure design: stored contracts, bounded tasks, reviewable handoffs, and one practitioner's account of keeping context across sessions.Field Journal · FJ-09When a Prompt Becomes a Small CharterA first-person proposal for contract-driven prompting: clearer inputs, boundaries, and return paths for consequential AI-assisted work.Field Journal · FJ-10The Hearthside Meta-ArchitectA personal reflection on building a humane, governed workbench for nonlinear systems work.Field Journal · FJ-11What I Mean by a Cognitive WorkbenchA working account of a cognitive workbench: a person-owned place where records, decisions, and carefully bounded tools help someone find their way back to the work.Field Journal · FJ-12When the Language Can Call a ModelA field note on HCAN's referent and effect vocabulary, Sigil's typed-delegation design, and an attributed historical report of a local model canary.
Fieldcraft methods
FieldcraftMake the acceptance condition executableTurn “looks right” into a condition you can inspect.FieldcraftTreat prompt fields as data, not separatorsA delimiter is part of the data eventually. Choose framing that survives the data you expect.FieldcraftCheck output shape before the next stepA result can be plausible text and still be unusable data. Validate the shape before consuming it.FieldcraftKeep a simple baseline in the comparisonA candidate is easier to interpret when the comparison target is visible and simple enough to audit.FieldcraftStop refinement at a declared budgetRefinement is useful until its cost and stopping rule disappear. Make the budget part of the trace.FieldcraftCarry the constraints across a handoffA handoff is a new boundary. Compare the receiving brief with the constraints that made the original task meaningful.FieldcraftScore a creative candidate with declared criteriaArithmetic can expose a brief fit or threshold while leaving aesthetic judgement with the human reviewer.FieldcraftVary one creative parameter at a timeIsolate a change so the comparison has a chance to teach you what changed.FieldcraftSeparate the fixed brief from optional variationGive exploration room by naming what cannot move and what may move.FieldcraftChoose a diverse concept set with declared tagsDiversity is not a mood word. Declare the tags and coverage floor you intend to preserve.FieldcraftCheck design contrast with the WCAG formulaPut the contrast calculation beside the palette decision so a readable design is testable before it is polished.FieldcraftDetect source revision drift with explicit identityCompare the source identity you expected with the identity you actually received.FieldcraftRequire every claim to name its evidenceMake support visible one claim at a time before synthesis hides the gaps.FieldcraftSeparate contradiction from missing evidenceA gap in support and a conflict in support require different next moves.FieldcraftPreserve minority conclusions during synthesisA synthesis can report agreement without erasing a material dissent.FieldcraftDetect cycles in a dependency graphA plan that depends on itself cannot be ordered by waiting harder.FieldcraftSelect context by relevance under a token budgetMake the selection rule visible when the context window cannot hold everything.FieldcraftPreserve source offsets when extracting spansRecord the index unit or a quotation can no longer be reliably re-located.FieldcraftExplain a revision with set differencesSay what entered and left a revision before narrating why it changed.FieldcraftSeparate evidence confidence from release statusA well-supported claim can still be a draft, and a released item is not a confidence score.FieldcraftReject stale retrieval evidence at an explicit boundaryFreshness is a comparison between a recorded retrieval time and a declared limit.FieldcraftFail closed when a guard state is unknownPreserve the current state when the condition for moving cannot be evaluated.FieldcraftKeep capability separate from permissionRequire the requested action and scope to match an explicit permission exactly.FieldcraftReturn a loss report with every projectionMake every dropped or normalized field visible to the next reader.FieldcraftCheck producer and consumer compatibilityCompare required field names and types before connecting two components.FieldcraftDetect idempotency-key conflictsBind a key to a request digest before deciding whether to replay or reject.FieldcraftDo not redispatch an indeterminate model stepPreserve the saved claim when continuity is lost and require a new recovery decision.FieldcraftCap deterministic retry backoffBound both the number of attempts and the exponential delay.FieldcraftEncode review state with Rust typestateMake release-candidate operations unavailable on draft values.FieldcraftMake parse failure explicit with Rust ResultReturn either a bounded integer or a typed reason the input was rejected.FieldcraftTransfer Rust ownership at a task boundaryConsume a task packet so the sender cannot keep using the same value.FieldcraftSeparate pure plans from released effectsLet planning describe work while effect admission checks exact releases separately.
Adjacent corridors
From the Sacred Commons & the Record
The Alignment You Already HaveThe ethics wing — charter, invocation, and the eight-section essay that began the public record.The Algorithm Has No Category for PrayerKali, copypasta, and the moderation of sacred repetition — the essay at the heart of the case.The Mystery SchoolA free-forever school of the science of the sacred — the commons in motion.The Extended Mind, ImplementedTEVV-mapped method paper — the governed scaffold as inspectable instrument.