Skip to content

docs: production roadmap & ecosystem architecture - #49

Merged
hyperpolymath merged 1 commit into
mainfrom
claude/project-scope-discussion-1dvjn8
Jul 9, 2026
Merged

hyperpolymath merged 1 commit into
mainfrom
claude/project-scope-discussion-1dvjn8

Conversation

@hyperpolymath

Copy link
Copy Markdown
Owner

What

Adds docs/PRODUCTION-ROADMAP.adoc — a planning artifact (requested) for taking plasma from the current pre-1.0 engine to production, and for the attendant estate architecture.

Contents

Part 1 — Production readiness (plasma itself)

  • Engine completeness: the reserved constructs (content facts + file-matches-pattern, release facts, consistent-with/decision registries, overlay modify/override, more actions)
  • Robustness: .plasma.toml policy pinning, .gitignore-aware walking, typed exit-code contract
  • Quality gates: real fuzzing (replacing the placeholder over parse_spdx_expr/load_policy_str), property tests, a CI determinism gate, MSRV + benches
  • Distribution: crates.io, signed binaries, real guix/nix, SBOM + cargo-deny, published JSON Schemas
  • The formal core (Catala/OCaml verified against engine-v0-design.adoc) as the 1.0 gate

Part 2 — Attendant architecture

  • Three stable interchange contracts (FactSet, Evaluation, SARIF) + a proposed signed run-record/attestation
  • The agent-honesty trio: plasma as deterministic oracle; did-you-actually-do-that (before/after facts diff + pass-findings); somethings-fishy (which obligations broke)
  • boj-server: a plasma cartridge (mirroring the existing ssg-mcp trigger) + a thin MCP server (plasma_check/facts/fix_plan)
  • Hypatia: --format logtalk renderer + plasma fix/check as autofix backend/gate
  • reposystem/contractiles: plasma as the evaluation engine behind Mustfile/Trustfile invariants; K9 Nickel → policy generation
  • verisimdb / feedback-o-tron / vexometer / proven at the edges

Part 3 — Phased plan: v0.4 (integration surfaces + trio contracts) → v0.5 (MCP + boj cartridge + logtalk) → v0.6 (governance) → v0.9 (hardening) → v1.0 (formal parity).

Part 4 — Cross-cutting standards: SemVer + schema versioning, determinism as a CI invariant, frozen interchange contracts, supply chain, observability.

Notes

Documentation only — no code changes; grounded in the repo's own declarations (boj-build.yml, integrations/*.a2ml, NEUROSYM.a2ml, .hypatia/, contractiles.adoc). Wire formats for the sibling repos are inferred from those declarations; they can be firmed up against the real interfaces if access to did-you-actually-do-that / somethings-fishy / boj-server is available.

🤖 Generated with Claude Code

https://claude.ai/code/session_01JSVSF9HhKUhDrEV4xkMEXq


Generated by Claude Code

Plans the path from the current pre-1.0 engine to production, and specifies
how plasma integrates with the estate: the agent-honesty trio
(somethings-fishy, did-you-actually-do-that), the boj-server cartridge/MCP
surface, the Hypatia neurosymbolic scanner (logtalk renderer + autofix
backend), and reposystem contractiles (policies as the evaluation engine).

Covers engine completeness, hardening/distribution, the formal core as the
1.0 gate, the three stable interchange contracts (FactSet, Evaluation,
SARIF) plus a proposed run-record/attestation, and a phased v0.3→v1.0 plan.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01JSVSF9HhKUhDrEV4xkMEXq
@hyperpolymath
hyperpolymath marked this pull request as ready for review July 9, 2026 23:50
@hyperpolymath
hyperpolymath merged commit b8df7bd into main Jul 9, 2026
22 checks passed
@hyperpolymath
hyperpolymath deleted the claude/project-scope-discussion-1dvjn8 branch July 9, 2026 23:51
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants