/mold

Ceremony scales to the job. The Bounds pass picks one of three tiers from references/tiers.md: Quick writes a validated mini-spec, Light runs only the dialogue modes the open forks need, and Full runs the whole Flow below. Saving a spec needs no extraction approval. Cook still needs an explicit user request. /cheese's tier-1 escalation enters mini-spec mode; see ## Agent-invoked mini-spec mode.

Flow

1. Bounds pass — open the Goal / Decided / Asking / [AGENT-DECIDED] ledger with the goal pinned as one sentence; only an explicit user fork changes it. Split it into G-n clauses (references/handshake.md § Goal coverage). Map every input's goals and non-goals as one [AGENT-DECIDED] line; ask the user only when the goal is genuinely unknown or a leverage trigger fires. Run the shape check, then announce the tier with its reason (references/tiers.md). Quick exits here to ## Agent-invoked mini-spec mode without a write-approval turn. Upgrade the tier whenever the evidence changes; never downgrade silently.
2. Route — choose the secondary mode from references/modes.md, announce it, and correct false premises first.
3. Dialogue — consequential forks are the user's to pick. A fork is consequential per the leverage line in ../age/references/voice.md. Every other fork is [AGENT-DECIDED]. Supply options, trade-offs, and evidence before you ask. Ground each critical claim through code, the Validate Cycle, or a Prototype Cycle. Resolve every contradiction. Render the decision map after three consecutive fork questions, or on request.
4. Sketch — For work across modules or with a new public interface, run references/shape-check.md. Bind identity and role nouns to code referents. Record the Placement block; no bodies.
5. Plan and validate — run the fresh-context fork-coherence taste test with python3 skills/mold/scripts/mold.pyz taste-test; a failure reopens only named forks, and the third failed verdict stops. The taste review guides the spec but is not a Cook execution reference. For bound curd work, dispatch a typed PlannerRequest, validate its PlannerResultWriterView (one retry, then save without runnable inputs), and normalize on the host. A clear single-curd task can instead use standalone Cook without a planner. Show semantic curds and waves before any Cook choice. See references/curdle.md § "Typed planner dispatch".
6. Readiness check — Before saving, run the coherence self-check and strict spec validation. Save a draft when unresolved parent work remains; record each hold in execution_holds:. Finalization refuses ready while an entry remains. User approval is not a prerequisite for writing. Before Cook, require an explicit user selection of scope and route. See references/handshake.md and references/early-curds.md.
7. Curdle — Resolve the durable spec path with python3 skills/mold/scripts/mold.pyz artifact-path specs <slug> and read path from its JSON reply. Write the validated draft and read it back. Retain typed plan evidence and unresolved holds. Write local issue drafts and durable decisions; publish follow-ups only after their own approval. A concrete bounded curd may use references/early-curds.md while Mold continues shaping its parent. The selected spec path can enter standalone Cook; bound curd execution also receives validated direct refs, never a pointer.
8. Offer Cook or keep shaping — after reconciliation, run python3 skills/mold/scripts/mold.pyz curd-count. A saved draft is not an execution request. If the user selects bound curd execution, run scripts/mold.pyz finalize and pass only ready direct inputs. A standalone clear task can enter Cook by spec path without machine finalization. Otherwise keep shaping or stop without a Cook command.

Portability: rules. Slash commands are host renderings, not the control model.

Follow-up candidates

Every non-goal and explicit dialogue deferral becomes a [FOLLOW-UP?] follow-up candidate. Dispose of the set before a runnable Cook handoff; details: references/handshake.md § Follow-up disposition.

Modes

Modes: references/modes.md. Evals: references/evals.md. Canvas: review-canvas.md.

Agent-invoked mini-spec mode

/cheese's tier-1 escalation calls /mold after the call site passes all Cook fast-path checks. The Quick tier enters the same mode without a write-approval turn. It produces a validated draft without design dialogue. Return the spec path. After the spec validates, show it and ask one direct Cook question. On explicit selection, dispatch /cook --spec <path> for clear standalone work; do not invent a pointer or require a planner. Append --hard when the user passed it.

The full coherence checklist does not run in this mode. The agent-introduced-scope check still runs implicitly. Every distinguishing noun must come from the user's input or tier-2 /culture//briesearch synthesis. Never add one silently.

Full procedure, the mini-spec schema, and the ## Provenance rules: references/mini-spec-mode.md.

Preferred tools and fallbacks

Call source-code search, read, and edit backends according to code-intelligence-routing.md. Shape checks use semantic caller search plus dependency context; procedure: references/shape-check.md.

Mold-specific tools beyond source-code routing:

The grounding record precedes the first structured question: the ledger holds a probe result — citations or hallouminate: absent — first. Mark each unsupported claim [?] until settled.

Sub-agent context gate

/mold owns the dialogue, contradictions, and save-approval state. Do not delegate these items. Delegate evidence-heavy code work to a fresh-context explorer and external research to a researcher. Shape uses an explorer digest as input. Record parent-context exploration as a degraded path. See references/context-budget.md for budgets and checkpoints.

Gate graph

python3 skills/mold/scripts/mold.pyz gate-graph --render dot|svg|png|mermaid renders one gate model. Image targets use Mermaid when Graphviz is unavailable. Tests keep gate nodes aligned with the handshake checklist. See references/gate-graph.md.

Spec provenance and Test Contracts

Every Mold-produced spec carries a provenance marker in frontmatter:

source: mold-handshake | agent-mini-spec | mold-curd-mini-spec

A spec has no disposition field. ## Test Contracts is optional: a spec may include it, and no gate requires it. Row-level rules: references/curdle.md § Test Contracts.

Fork taste gate

python3 skills/mold/scripts/mold.pyz taste-test binds the verdict to draft SHA256 and each settled fork. Stale, partial, or blocked verdicts fail; a failure reopens only named forks, with two rounds. Approved specs pass unchanged metadata to /cook --spec <spec-path> (add --auto only when the user chose it) for standalone work; bound curd work uses /cook --auto --result <result-path>.

Each fork appears in Approach, Interface sketches, and Acceptance, plus Test Contracts when the draft has them. Do not rename sections. Tag reflecting lines with fork id and run taste-test --precheck before dispatch. goal must survive verbatim in Problem statement (goal-drift); each G-n clause carries an Acceptance or disposition tag (goal-coverage). See references/curdle.md § Spec template and references/gate-graph.md.

Execution gate

Mold may save a validated parent spec or early curd mini-spec without an approval turn. The agent coherence check still controls what the artifact claims as settled. Print the narrowing delta (taste-test --coverage) and the scope audit table before presenting any Cook option. Unresolved decisions remain visible holds, not implicit approvals. See references/handshake.md.

Never start Cook from a saved spec, silence, --auto, or a general design approval. The user must select the exact curd or parent scope and a Cook route. A changed scope or plan needs a new selection. A failed bound finalization leaves a saved draft without executable refs. See references/early-curds.md.

--hard

Mold never runs the metacognitive check. Plate alone runs it at the verified-artifacts boundary. Mold appends --hard to every Cook command that it emits when the user passed the flag. Cure forwards the same flag to Plate. See ../hard-cheese/references/composition.md.

Handoff

Pipeline: culture → [mold] → cook (⇄ press) → age → cure → plate

After Curdle's phase two finishes, run python3 skills/mold/scripts/mold.pyz curd-count for the blast-radius digest. A draft offers no Cook command. On explicit Cook selection, retain the selected spec path for standalone work. Bound curd work dispatches the printed finalize command, /cook --auto --result <result-path>; the result carries request_id, spec_ref, planner_result_ref, plan_ref, and exact coverage. Prompt through the shared handoff gate (policy); never pre-select. Keep applicability, contract, and taste review metadata unchanged. Append --hard when the user passed it.

The digest's mode is orientation, not a skill or Cook gate. Render the fixed blast-radius menu from decomposable, candidate_curds, verdict, and mode; see references/handoff-menus.md.

Rules

• Dialogue first; artifacts are the by-product.
• Tiered lettered options. Consequential forks use A/B/C/D choices via the question transport at ../cheese/references/ask-user-question.md. Never decide them silently. Everything below the leverage line is [AGENT-DECIDED] by default: make the call, log a one-line vetoable alternative in the ledger, and do not ask. A fork is valid only after its depth was contributed in-dialogue first. Precede every structured question with visible prose that weighs the fork and the evidence. Keep one open picker in conversation. In the optional review canvas, independent, already-explained fork cards may share one batch; dependent forks wait for agent reconciliation.
• Fork cards. Before each consequential question, show a fork card: what it decides, what it moves, and each branch as a symbol table or plain-text graph. Offer Explain more as a choice; an unclear reply re-renders the same fork with more context (fork-questions).
• Altitude tag. Every Asking fork names the acceptance criterion, public seam, or non-goal it moves. A fork that moves none, or that sits below the leverage line, is [AGENT-DECIDED] or a follow-up candidate, never a user question.
• Decision ledger. Each round prints Goal / Decided / Asking / [AGENT-DECIDED], the goal verbatim. Curdle persists consequential decisions to ADRs and minor ones to the spec. The taste verdict names every settled consequential entry exactly once.
• Decision map and fork-round cap. Stop after three consecutive fork rounds that add no new evidence, or three consecutive forks that fail the altitude tag, or when the user requests a decision map. Show completed forks, remaining forks, and a ready or blocked verdict. A fourth round requires new grounding, a delegated digest, or a /wheypoint checkpoint. The map shows ledger state. It does not create an artifact.
• Do not implement code.
• Do not write production files in Mold. Saving a spec does not authorize Cook.
• Do not silently settle uncertain claims.
• Apply the shared voice kernel at ../age/references/voice.md. Correct false premises. Mark each critical claim's confidence as certain | speculating | don't know. Steelman before you dismiss. Put consequential forks to the user; depth informs the question, never replaces it.

The generated ../cheese/references/schema-intertwine.md lists every published schema and its contract model.

Agent resolution

Resolve delegates through ../cheese/references/agent-resolution.md.

The canonical mold spec or mini-spec carries the shared agent_resolution block. Commands: references/commands.md.