Roadmap

Pegma ships component by component, each through its own phased plan, each gated by real consumers rather than dates. This page is the aggregate view.

Status lines below are compiled at build time (2026-11-16) from each repository's public docs/PROJECT_PLAN.md — 14 of 16 components carry a plan; the rest (and any fetch miss) fall back to a hand-written snapshot dated 2026-08-15. The repos win any disagreement.

Where things stand

Component Status Now
Spine published 0.1.2 on npm; deliberately close to frozen.
Storage Core published Fallback snapshot (2026-08-15): 0.4.0 on npm with the Azure Tables (0.4.0) and Cloudflare D1 (0.4.1) adapters; both adapters pass the same conformance suite.
Storage Blobs published Phase 6 complete. @pegma/storage-blobs@0.2.0 and the Azure Blob, Cloudflare R2, and S3 adapters publish at the same version with trusted-publisher provenance from signed annotated release tags. 0.2.0 adds an optional cacheControl on put, stored as native object state by every first-class backend and returned by head/get.
Cache Core published Phase 5 — Upstash Redis adapter, published 0.1.1. @pegma/cache-core, @pegma/cache-conformance, @pegma/cache-redis, @pegma/cache-azure-redis, @pegma/cache-elasticache, and @pegma/cache-upstash-redis are published at 0.1.1. The 0.x API is unstable.
Authorization Core published Phase 5 complete; Phase 6 integration feedback and stabilization next (the signed v0.1.0 release published the original seven packages through OIDC with provenance on 2026-07-27; synchronized v0.1.2 corrects the pre-publish v0.1.1 release-test failure and first advertises Identity as the eighth package; synchronized v0.1.3 resolves the 2026-07-29 security review; synchronized v0.1.4 first advertises Entra as the ninth package; synchronized v0.2.0 aligns @pegma/storage-core at 0.4.0 across the stack; synchronized v0.3.0 delivers the identity-link write and lifecycle listing from the first reference-consumer feedback; synchronized v0.4.0 is prepared, first advertising the Admin role-administration service as the tenth package; 0.x public API remains unstable)
Audit published Phase 1 complete; awaiting the first real consumer (0.1.1 published, public API unstable; 0.2.0 is prepared, aligning @pegma/storage-core at 0.4.0 across the stack)
Support Desk published Buildout Tasks 1–6 are complete; the first 0.1.0 set was published on 2026-07-29 with signed tag v0.1.0. Later Task 11 package work, Tasks 13–14, and SD-01 through SD-06 are merged source. The standalone KB has merged KB-01/02/03 and independent partial KB-05 preparation, KB-04's transport source exit, and KB-05 release preparation. AUTH-01 Agent Access source and its separate release preparation are also merged. The ten entries in the historical source ledger record the exact PRs/commits, verified gates, and remaining dependency order.
Webhooks published Phases 1 through 4 are complete. @pegma/webhooks@0.1.1 is published from the protected signed v0.1.1 release through npm trusted-publisher OIDC with provenance (0.x, public API unstable)
Sessions published Phases 1–4 are complete. The published 0.2.0 release aligned @pegma/storage-core at 0.4.0. RetireGolden is the first integrated consumer and pegma.dev is the second, live consumer. The 0.2.1 patch moves default clock observations after awaited reads and sweep snapshots. Publication is recorded in GitHub releases and npm; source metadata alone does not publish a version. The public API remains unstable.
Mail published Phases 1–3 are complete. @pegma/mail@0.1.1 is published and is composed by both Identity and Support Desk. The public API remains unstable.
Billing Core published Phase 4 — Stripe adapter in-tree. @pegma/billing-core and @pegma/billing-stripe are published at 0.1.1. The 0.x API is unstable. Phase 5 stays soak-gated.
Identity published Phases 1–4 are complete. @pegma/identity@0.1.1 is published and live in pegma.dev with @pegma/authorization-identity@0.1.2. The public API remains unstable. 0.1.2 is prepared, aligning @pegma/rate-limit at 0.2.0 and @pegma/mail at 0.1.1 so the dependency tree resolves a single @pegma/storage-core@0.4.0. 0.1.1 was a documentation-only patch over 0.1.0 that closed the operator-guidance findings from the 2026-07-28 security scan; it changed no runtime behavior and no public API.
Flags Core published Published 0.1.1 (2026-08-15). Public API unstable (0.x).
Logger Adapters published 0.1.3 on npm (tee, Application Insights, Cloudflare, Datadog); trusted publishing configured; OIDC publish workflow in-repo. Public API unstable. Host tee wiring on RetireGolden and pegma.dev is the remaining consumer bar.
Health published 0.1.2 on npm; trusted publishing configured. 0.2.0 is prepared, aligning @pegma/storage-core at 0.4.0 across the stack. Public API unstable.
Rate Limit in development Phases 1, 2, and 4 are complete; Phase 3 has its first production consumer. pegma.dev composes the durable tier for its Identity surface. Registry latest is published 0.3.2; exact-byte and npm cryptographic checks passed, but strict postflight rejected its authenticated builder identity. It is not qualified for the planned helper adoption. Current source is the unpublished 0.3.3 candidate, retaining exact Storage Core 0.4.0. Published 0.2.0 predates the already-merged PR #8 hardening for immutable policy descriptors, counter kind/configuration bindings, and clock-skew grace. The candidate includes that hardening plus bounded global maintenance. The public API remains unstable; source completion is not a publication or host-adoption receipt.

What gates what

The dependencies that actually pace the work — none of them are dates:

What 1.0 means

Nothing reaches 1.0 on ambition. The ecosystem rule is that breaking changes stay permitted until multiple real consumers complete adoption — a component earns stability by being used, not by being finished. Until then, every package is honestly 0.x.