Charge — M0 Platform charter

Sprint 0 goal

Lock charter-grade platform intent: bounded product responsibilities, intersecting contracts, readable roadmap and sprint scaffolding, validated JSON samples, and adoption hooks (boundary stubs in product repos). A reviewer can onboard from README.md Start here through platform-charter.md → product-boundaries.md → schemas without guessing ownership.

Product owners (assign names)

Active stories

• M0-A — Freeze ADR summaries (already drafted) against docs/intersection-map.md; flag any contradiction for a Spark or ADR follow-up instead of silent drift.
• M0-B — Ensure docs/forge/product-boundary.md exists in forge-lenses, forge-lcdl, forge-fleet (thin stubs from platform product-repo-stubs/ copied via scripts/sync_product_repo_stubs.py); reconcile wording with canonical docs/product-boundaries.md on first edit.
• M0-C — Run prompts/shared/platform-review-assay.md as platform assay (Composer 02) over charter, boundaries, roadmap, sprint structure, and schemas/forge_run.v1.schema.json + schemas/evidence_packet.v1.schema.json; record pass / warning / fail in evidence/platform-assay-log.md.
• M0-D — Capture scripts/check_foundation.py CLI success in evidence/foundation-check.log.md or paste output into assay log appendix.

Active Sparks (next backlog pulls)

1. SPK-traceability — Map one worked example: ForgeRun sample ↔ EvidencePacket ↔ Lenses run language (minimal UI glossary note in docs/glossary.md if gaps appear).
2. SPK-approval-minima — List minimum human gates per approval_mode for proposal and sandboxed_implementation; align Fleet template exposure with ADR-0003.

Human gates

• Human sign-off that Lenses-as-system-of-record (charter + ADR-0001) matches how Fleet/LCDL surface state today or documents the gap intentionally.
• Cross-product acknowledgment of Agents as workcells (ADR-0002) across any operator docs that imply autonomous final state.
• Review of synced docs/forge/product-boundary.md stubs in sibling repos before merge there (ownership + escalation path).

Evidence required

• evidence/platform-assay-log.md with assay outcome + boundary-risk bullets.
• evidence/foundation-check.log.md (or pasted transcript) proving green python3 scripts/check_foundation.py run.
• Links to sibling PRs that land boundary stubs (forge-lenses, forge-lcdl, forge-fleet).

Acceptance criteria

• Boundaries readable in narrative + ADR form without contradictions flagged fail by assay rubric (warnings allowed with mitigation owners).
• check_foundation passes from repo root post-change.
• All three sibling repos contain non-empty docs/forge/product-boundary.md aligned to product-repo-stubs/*/docs/forge/product-boundary.md or explicitly extended with product-specific escalation.

Risks

• Divergent understanding of approval_mode semantics between Fleet launch paths and LCDL prompts.
• Boundary stubs stale if product repos refactor docs/ layout — document redirect or move plan per repo.

Expected demo

Walk Start here in README.md, open docs/platform-charter.md and docs/intersection-map.md, sketch one ForgeRun with linked EvidencePacket from samples, open a sibling docs/forge/product-boundary.md stub in review.