Cross-format data model + milestone & quality SOWs#1
Merged
Conversation
…roject split into milestones ## Cross-format data model analysis (SOW-0002) Evidence-based investigation of all five source formats: ai-agent v2/v3, claude-code, codex, opencode. Real fixtures inspected (294,316 v2 + 17,356 v3 + 3,614 claude-code + 2,585 codex rollouts + 3.9 GB opencode SQLite), upstream source code cited at file:line. Five per-client adapter specs rewritten from sketches (3,088 lines total). Findings absorbed into canonical-events.md and data-model.md as first-class fields rather than extras_json bloat: - Op kinds: added `system` (v2/v3) and `compaction` (claude-code/codex) - Session statuses: added `abandoned`, `interrupted`; documented indefinite-`running` for claude-code (no terminal signal) - Session kinds: added `fork` (codex forked_from_id) - Cache tokens (`tokens_cache_read`, `tokens_cache_write`) on sessions, turns, ops — surfaced by v3, claude-code, opencode - `provider_alias` for opencode's user-defined provider names - `reasoning_kind` (`summary` | `raw`) for codex's reasoning split - `chars_in/out` for v2's character-based tool accounting - `cwd`, `call_path`, `last_activity_ts` promoted from extras_json to first-class columns - `PayloadRef.sha256` - `RootNativeID` and `ParentOpKey` on SessionStartedEvent - `Status='truncated'` on ops for codex limited-mode payloads - `catalog_providers`, `catalog_cwds` rollup tables Spec adds explicit cross-format compatibility matrix and sub-agent linkage strategy table. ## SOW-0001 moved to current/ with chunk renumber Status: open -> in-progress. Implementation plan renumbered to put "Spec deltas" as Chunk 1 (UI theme, pricing.md, adapter spec refinements); CI scaffolding becomes Chunk 2. Honors the spec-first-then-test-then-code discipline. ## Milestone SOWs (pending operator approval, M2-M5) - SOW-0003 claude-code adapter - SOW-0004 codex adapter - SOW-0005 opencode adapter - SOW-0006 APM-style tracing UI (topology + timeline + span view) - SOW-0007 Statistics & analytics (rollups + FTS5 search + dashboard) - SOW-0008 Release polish + v0.1 tag ## Quality automation SOWs (pending operator approval) - SOW-0009 Go static analysis stack (golangci-lint, gosec, govulncheck) - SOW-0010 Test infrastructure + coverage gates - SOW-0011 Fuzz, property-based, benchmark infrastructure - SOW-0012 Frontend quality stack (ESLint flat, strict TS, Vitest coverage, Playwright + axe, bundle budget) - SOW-0013 Repo-wide gates + canonical CI workflow Each quality SOW cites the specific gates from quality-gates.md it lands. Together they enforce every threshold in the spec.
7 tasks
ktsaou
added a commit
that referenced
this pull request
May 30, 2026
…ctivate Prepare the opencode adapter SOW (the last of the five). opencode is SQLite-backed (~/.local/share/opencode/opencode.db, WAL, ~4.36 GB live), so the adapter replaces the JSONL parser+scanner+stream+tailer with a read-only SQL delta-query layer + a watermark cursor + a poll loop. Resolve two spec TBDs before any test/code: ratify task->session "emit both" (tool Op + session Op as topology parent); firm the per-turn token rule to delta-from-previous-message-totals with an explicit implementer-verify-on-live-DB note (only the step-finish cumulative pattern is independently verified so far). Fill the gate (problem model, live-DB evidence, SQLite structural map, watermark cursor, canonical mapping incl. the verified cumulative-token delta, read-safety against the live concurrent writer, sensitive-data plan, chunked implementation + validation plans tied to acceptance #1-8, and six recorded CTO decisions) and move the SOW to current/. Purely additive: a new internal/adapters/opencode/ package + an additive auto-discovery probe in cmd/ai-viewer-ingest/sources.go; no canonical, ingest, or store change (the catalog is already idempotent post-SOW-0004 and every target canonical field exists).
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
This PR lands two deliverables operator requested on 2026-05-26:
quality-gates.md.SOW-0001 also moves to
current/with chunk renumber (specs first per the discipline contract).What's in this PR
Specs updated (cross-format unified model)
.agents/sow/specs/canonical-events.md— full rewrite. AddsOpKindvaluessystem+compaction;SessionStatusaddsabandoned/interrupted; cache tokens; provider_alias; reasoning_kind; chars_in/out; cwd; call_path; RootNativeID/ParentOpKey; cross-format mapping table..agents/sow/specs/data-model.md— full rewrite. Schema reflects every new field; newcatalog_providers+catalog_cwdstables; explicit cross-format compatibility matrix; sub-agent linkage strategy table..agents/sow/specs/adapter-aiagent-v3.md— evidence-based rewrite (719 lines)..agents/sow/specs/adapter-aiagent-v2.md— evidence-based rewrite (456 lines)..agents/sow/specs/adapter-claude-code.md— evidence-based rewrite (795 lines)..agents/sow/specs/adapter-codex.md— evidence-based rewrite (551 lines; openai/codex@8a94430b)..agents/sow/specs/adapter-opencode.md— evidence-based rewrite (567 lines; anomalyco/opencode@2b3ddf9f).SOWs
pending/SOW-0001-phase-1-foundation.md→current/. Status: open → in-progress. Implementation plan renumbered: Chunk 1 = spec deltas, Chunk 2 = CI scaffolding (was Chunk 1).SOW-0002-20260526-cross-format-data-model-analysis.md— records the analysis work.SOW-0003-20260526-claude-code-adapter.mdSOW-0004-20260526-codex-adapter.mdSOW-0005-20260526-opencode-adapter.mdSOW-0006-20260526-apm-tracing-ui.mdSOW-0007-20260526-statistics-and-analytics.mdSOW-0008-20260526-release-polish-v0-1.mdSOW-0009-20260526-go-static-analysis-stack.mdSOW-0010-20260526-test-infrastructure-coverage.mdSOW-0011-20260526-fuzz-property-benchmark.mdSOW-0012-20260526-frontend-quality-stack.mdSOW-0013-20260526-repo-wide-gates-ci-workflow.mdInvestigation methodology
Five parallel deep-research subagents (one per format), each instructed to:
Master assistant synthesized the gaps into the canonical updates.
Stats
Test plan
quality-gates.md.