Skip to content

ROADMAP.md

Latest commit

 

History

History
258 lines (203 loc) · 13.8 KB

ROADMAP.md

File metadata and controls

258 lines (203 loc) · 13.8 KB

Roadmap

Public roadmap for opencode-sdlc-wizard. Items are ordered by current priority but may shuffle as real-world use surfaces friction.

v0.2.0 — shipped 2026-05-04

  • ✅ Privacy-first four-tier backend picker (scripts/{detect,configure}-backend.sh)
  • ✅ Live E2E reliability fixes (session.created race + Bun.spawnSync)
  • ✅ Codex round-3 cross-model review CERTIFIED 9/10
  • ✅ 113/113 tests across 4 suites
  • ✅ npm publish flow + release.yml automation
  • ✅ Companion docs: PRIVACY.md, RELEASING.md, HANDOFF.md addendum

See CHANGELOG.md for the full release notes.

v0.3.0 — npx opencode-sdlc-wizard init CLI — shipped 2026-05-04

  • ✅ bin entry in package.json → cli/bin/opencode-sdlc-wizard.js
  • ✅ Node wrapper that shells out to install.sh for the real run
  • ✅ Wrapper-level --dry-run previews bundle without writing
  • ✅ --target-dir / --force pass through to install.sh
  • ✅ 10 CLI tests in tests/test-cli.sh (123/123 total across 5 suites)
  • ✅ README updated to lead with npx opencode-sdlc-wizard init

v0.3.1 — cross-model-review skill (OSS reviewer) — shipped 2026-05-04

  • ✅ skills/cross-model-review/SKILL.md — adaptive skill picks reviewer + runs through opencode
  • ✅ scripts/cross-model-review.sh — non-interactive wrapper, alias-aware
  • ✅ Provider alias resolution shared with configure-backend.sh
  • ✅ 10 tests via stubbed opencode (138 total across 6 suites)
  • ✅ CHANGELOG + skill recommendations documented

Live E2E pending: needs a configured OSS provider (Groq free tier, local Ollama 30B+ model, or Together/OpenRouter key). Stub tests cover the script logic; live E2E proves the wiring against a real reviewer.

v0.4.0 — domain-adaptive TESTING.md templates — shipped 2026-05-04

  • ✅ 4 templates: firmware / data-science / cli / web
  • ✅ Setup-wizard skill detects domain from concrete signals
  • ✅ Templates ship at .opencode/templates/testing/<domain>.md
  • ✅ 26 tests in test-domain-templates.sh (164 total across 7 suites)

v0.5.0 — check subcommand + bundle-drift extension — shipped 2026-05-05

  • ✅ npx opencode-sdlc-wizard check [--json] [--target-dir] — symmetric to init; rc=0 (current) / rc=1 (behind) / rc=2 (not installed)
  • ✅ scripts/check-updates.sh — backs the CLI; also installable at .opencode/scripts/ so skills + hooks can invoke directly
  • ✅ Fixed hook bug: npm view agentic-sdlc-wizard → opencode-sdlc-wizard
  • ✅ Extended drift tests to cover hooks (no parent-package executable refs, mirror byte-equality)
  • ✅ 221 tests across 9 suites (was 192/8 in v0.4.1)

v0.6.0 — SDLC.md + ARCHITECTURE.md templates — shipped 2026-05-05

  • ✅ templates/sdlc.md — SDLC baseline + workflow phases + confidence levels + cross-cutting rules; references both codex + cross-model-review
  • ✅ templates/architecture.md — overview / components / environments / deployment / decisions log
  • ✅ install delivers both at .opencode/templates/
  • ✅ setup-wizard skill references both templates
  • ✅ 17 tests in test-doc-templates.sh (238 total across 10 suites)

v0.7.0 — JSON schemas for review artifacts + zero-dep validator — shipped 2026-05-05

  • ✅ templates/schemas/handoff.schema.json — draft-07 schema for the handoff artifact (review_id / status / round / mission / success / failure / review_instructions + optional fields)
  • ✅ templates/schemas/response.schema.json — draft-07 schema for the response artifact, including conditional validation (FIXED requires fix_summary+fix_locations; REJECTED requires rejection_reason) and patternProperty support for recheck_instructions_for_round_N
  • ✅ scripts/validate-review-artifact.sh + .js — zero-dep node validator (no npm install to consumers); handles draft-07 subset the schemas use including $ref + allOf if/then + const
  • ✅ install lands schemas at .opencode/schemas/ + validator at .opencode/scripts/
  • ✅ cross-model-review skill new Step 1.5 — validate before sending the prompt; setup-wizard skill mentions schemas in Step 4
  • ✅ Drift-test extension: scripts/.js + templates/schemas/ coverage
  • ✅ 28 tests in test-review-schemas.sh (270 total across 11 suites)

v0.8.0 — free-tier-first cascade + 5 providers + cost ladder — shipped 2026-05-05

  • ✅ --free-tier-first flag on detect-backends.sh (and DETECT_FREE_TIER_FIRST=1 env) biases cascade toward free providers
  • ✅ 5 new providers wired through detector + configure-backend: Cerebras, DeepSeek direct, NVIDIA NIM, Google AI Studio (Gemini), MLX (Apple Silicon native)
  • ✅ docs/cost-ladder.md — concrete $0 / $20 / $200 monthly budget paths with per-job picker table
  • ✅ README bumped + cross-links to cost ladder
  • ✅ package.json files[] adds docs/ so cost ladder ships in tarball
  • ✅ 8 new picker tests + 7 new doc-template tests (285 total / 11 suites)

v0.8.4–v0.8.9 — drift family closure — shipped 2026-05-12

Cleanup sweep after v0.8.0 added 5 new picker providers + the cost ladder doc. Five surfaces drifted past the v0.8.0 picker; each was closed with a regression gate.

  • ✅ v0.8.4 — install.sh "next steps" hint refreshed for v0.8.x picker
  • ✅ v0.8.5 — validator emits DOMAIN HINT on foreign-domain artifacts
  • ✅ v0.8.6 — sibling-wizard awareness (.claude/, .codex/) + setup-wizard SKILL.md drift
  • ✅ v0.8.7 — cross-model-review SKILL.md drift + stale DeepSeek price (~$0.27/M → ~$0.14/M)
  • ✅ v0.8.8 — AGENTS.md + PRIVACY.md tier-table + walkthrough drift
  • ✅ v0.8.9 — cross-model-review.sh wrapper alias parity with configure-backend.sh (only runtime bug in the family — nvidia_nim/google_aistudio silently broke; caught by codex pre-ship review)
  • ✅ 305 tests across 11 suites

v0.9.0 — pick subcommand (one-shot detect → configure) — shipped 2026-05-12

Codex's direction call: collapse picker UX before Mixed-mode. The two-step detect → configure workflow becomes npx opencode-sdlc-wizard pick [--free-tier-first] [--tier T] [--provider P].

  • ✅ scripts/pick-backend.sh — orchestrator with PATH-first script lookup
  • ✅ Canonical default-model map per tier/provider (14 entries — single source of truth)
  • ✅ CLI subcommand wired in cli/bin/opencode-sdlc-wizard.js
  • ✅ install.sh ships pick at .opencode/scripts/pick-backend.sh
  • ✅ 25 new tests in test-pick.sh (333 total / 12 suites)
  • ✅ Default-model map drift gate (T12) — every v0.8.x picker provider must have a default

v0.9.1 — Ollama default Qwen 2.5 → Qwen 3 — shipped 2026-05-12

Community-patterns research surfaced that Qwen3-Coder is the most-cited local default in May-2026 shared configs. Single default-model bump verified against the live ollama.com tag list before shipping.

  • ✅ private_local/ollama default: qwen2.5-coder:32b → qwen3-coder:30b
  • ✅ All mirroring doc surfaces updated (install.sh, PRIVACY.md, README, AGENTS.md, cost-ladder)
  • ✅ Other community-flagged swaps (Groq, OpenAI, Gemini, DeepSeek) deferred per-provider until each lives in a verifiable catalog

v0.10.0 — Mixed-Mode (per-agent model routing) — shipped 2026-05-17

The pattern community research called out as the May-2026 baseline: 11/15 surveyed opencode.json files route review work to a different model than build work. v0.10.0 makes the split a one-flag-pair operation.

  • ✅ configure-backend.sh learns --reviewer-tier T --reviewer-provider P --reviewer-model M (all-or-nothing triplet)
  • ✅ Writes agent.review.model plus reviewer's provider block (deep-merged with coder's; same-provider collapses to one block)
  • ✅ Reviewer side uses same PROVIDER_ALIASES map → canonical IDs only in the written config
  • ✅ pick forwards reviewer flags; --reviewer-model optional (filled from canonical default-model map, same lookup as coder)
  • ✅ Default-model lookup factored into a default_model_for() function — single source of truth used for both sides
  • ✅ 8 new tests (T31–T34 in test-backend-picker, T15–T18 in test-pick); 341 total across 12 suites
  • ✅ docs/cost-ladder.md $20/mo hybrid example shows the new one-shot invocation
  • ✅ cross-model-review.sh wrapper unchanged — Mixed-Mode is the standing config; wrapper is the explicit invocation

Planner / docs / test-writer agent routing deferred to v0.10.1+ — pick gains --planner-* / --docs-* flags as each pattern is validated.

v0.10.1 → v0.13.1 — community-signal sprint — shipped 2026-05-17 / 2026-05-18

Once pick existed, the May-17 community-patterns research surfaced a clear queue: 5/6 of the highest-signal community patterns ranked above 50% adoption in surveyed opencode.json files. The v0.10.1–v0.13.1 arc shipped every one of them. Wizard is now feature-complete relative to May-2026 community signals.

  • ✅ v0.10.1 — --sandbox-test-writer + --sandbox-docs (path-scoped permission blocks)
  • ✅ v0.10.2 — --planner-* (Mixed-Mode for the plan agent; community's #1 most-overridden agent at 57%)
  • ✅ v0.10.2 — default-model bumps: openai → gpt-5.3-codex, google → gemini-3.1-pro
  • ✅ v0.10.3 — default-model bumps: deepseek → deepseek-v4-flash, groq → gpt-oss-120b
  • ✅ v0.10.4 — --small-* for top-level small_model (35–40% community adoption)
  • ✅ v0.10.5 — --sandbox-plan (categorical tool denial via agent.plan.tools.{write,edit,patch}=false)
  • ✅ v0.10.6 — cost-ladder.md recalibration (model IDs + Z.AI quarterly pricing)
  • ✅ v0.11.0 — Z.AI GLM Coding Plan as proprietary provider (post-Anthropic-OAuth-ban migration target)
  • ✅ v0.11.1 — --coder-temp / --planner-temp / --reviewer-temp per-agent temperatures
  • ✅ v0.11.2 — full security agent: --security-* triplet + --security-temp + --sandbox-security
  • ✅ v0.12.0 — new managed tier; OpenCode Zen (opencode provider; 40+ models incl. free tier)
  • ✅ v0.13.0 — new subscription tier; GitHub Copilot Pro+ (first OAuth-based provider, no env-var)
  • ✅ v0.13.1 — PRIVACY.md tier walkthroughs updated for the six-tier reality

Wizard surface as of v0.13.1:

  • 6 tiers: private_local, enterprise, hosted_oss, managed, proprietary, subscription
  • 17 providers with canonical default-model entries
  • 5 first-class agents: coder/build, small_model, planner (plan), reviewer (review), security
  • 3 sandbox shapes: path-scoped permission.write (test-writer, docs), categorical tools-denial (plan, security)
  • 407 tests across 12 suites
  • 15 tagged releases this sprint (v0.8.4 → v0.13.1)

v0.13.2+ candidates (post-sprint backlog, no community-research backing)

These are speculative — they have no direct community-signal evidence behind them. Either dogfood feedback or a fresh research pass should drive what's next.

  • Auto-nudge integration: instructions-loaded-check.sh hook delegates to check-updates.sh instead of duplicating the version- check logic. DRY refactor; drift prevention only.
  • NIM 1M-context emphasis in cost-ladder: per May-2026 research, NVIDIA NIM free tier has 1M-token context on DeepSeek V4 with RPM-only limits — bigger than Cerebras's 8K-64K free cap. Doc patch.
  • OPENCODE_SDLC_WIZARD.md master doc: equivalent of parent's 4506-line CLAUDE_CODE_SDLC_WIZARD.md. Heavier lift; defer until consumer feedback says it's needed.
  • Schema versioning + migration: when v0.7.0 schemas need a breaking change, add $schema_version field + a migrator the validator runs through. Defer until first breaking change is needed.

Phase B — backend matrix proof

Status: deferred from v0.1.0. Triggered by maintainer authorization.

Run paired E2E SDLC scenarios (plan → TDD → self-review) against:

Tier Backend Floor model
Local Ollama / Qwen2.5-Coder-32B (16-24GB VRAM class) Floor
Enterprise Azure OpenAI tenant Floor
Hosted OSS Together / Groq / OpenRouter (Qwen-Coder, DeepSeek-V3) Floor
Anthropic baseline Opus 4.7 max Ceiling

Score each against parent's tier-2 evaluation rubric (10-point criteria, 5 trials, 95% CI). Document which backends hold SDLC compliance and which degrade. Capability-floor note still applies: small models (7-13B) will fail and that's a result, not a bug.

Phase C — hardware scout for local tier

Status: deferred. Triggered if Phase B needs hardware.

Test gaming laptop + Windows laptop first (zero spend). If 16-24GB VRAM bar isn't met, evaluate $200/$300/$400 rig OR cloud-GPU rental.

Cross-cutting — XDLC ditto migrator

Status: roadmapped, not in this repo's scope.

A separate cross-host migrator (ditto) is roadmapped at xdlc/docs/ditto-roadmap.md to automate future ports between siblings. v0.1.0 trigger: when sibling #5 starts (or upstream claude-sdlc-wizard ships a release that needs syncing across the three existing siblings).

The first transform set ditto v0.1.0 will codify is the claude-sdlc-wizard → opencode-sdlc-wizard port we hand-applied for v0.2.0 — captured in HANDOFF.md v0.2.0 addendum + the F1-F6 + E2E findings tables in .reviews/response.json. Those are effectively the unstructured form of xdlc/docs/ports/claude-to-opencode.md.

When ditto v0.1.0 ships, its first regression test will harvest those tables, scan + diff the two repos, and assert the auto-generated transform table matches what we wrote down by hand.

How to consume this roadmap

  • Open issues in the repo to bid up an item or propose a new one.
  • Items are not promises. Real-world use of v0.2.0 may surface higher-priority work that bumps these.
  • feedback skill (privacy-first) is the canonical way to surface bugs / requests without scanning the repo.