Skip to main content

Cross-project capability map

This map groups capabilities by what the software can do, not by repository.

Maturity uses the Product Compass reuse model: local implementation, repeated pattern, documented pattern, extracted component, governed package.

Capability groups

Data ingestion and monitored pipelines

CapabilitySummarySource projectsMaturityReuse potentialLikely extraction shapeProducts it could helpEvidence status
Monitored source ingestionDetect source changes, fingerprint inputs, create/resume runs, store artefacts, load warehouse tables, validate, and finalise.DataStitcher, SiteIntel bulk ingest, BigQuery AtlasDocumented patternHighPattern first, then works-ingest contracts/packageDataStitcher, SiteIntel, BigQuery Atlas, Bollocks Radar, future public-data productsFact plus inference
Resumable run and checkpoint stateStore run steps, attempts, artefact pointers, retry state, and failed-stage recovery.DataStitcher, SiteIntel bulk ingest, BigQuery Atlas cataloguerRepeated patternHighSchema pattern, TypeScript contracts, admin UI patternAny product doing scheduled data collectionFact
Source registryRepresent sources, cadence, enabled state, credentials required, table specs, parser type, and downstream outputs.DataStitcher, SiteIntel bulk ingest, Trackpal provider registryRepeated patternHighShared source/provider registry contractTrackpal, SiteIntel, DataStitcher, BigQuery AtlasFact plus inference
Raw artefact boundaryStore raw files/payloads in object storage and keep databases limited to summaries, state, and pointers.DataStitcher, SiteIntel, Bollocks Radar, BigQuery AtlasDocumented patternHighGuide plus validation rulesEvery ingestion productFact

Crawling, discovery, and enrichment

CapabilitySummarySource projectsMaturityReuse potentialLikely extraction shapeProducts it could helpEvidence status
Evidence-backed candidate discoveryFind candidates only from public evidence, score provenance, dedupe, and cool down repeat attempts.BigQuery Atlas, DataStitcher public audit, SiteIntel bulk ingestDocumented patternHighPattern plus candidate/evidence schemaBigQuery Atlas, DataStitcher, SiteIntel, public-good toolsFact
Domain intelligence task registrySplit domain collection into source-specific tasks with TTLs, dependencies, queues, retries, cost tiers, and worker targets.SiteIntelLocal implementation / documented patternHighShared task registry contract after second consumerSiteIntel, Trackpal website monitor, future smoke-alarm toolFact plus opportunity
Website map analysisAnalyse robots, sitemaps, homepage metadata, schema, auxiliary files, platform hints, and display-friendly insights.SiteIntel ideas and current task docsPrototype / local implementationMediumStandalone tool or source adapterSiteIntel, free website smoke-alarm tool, Trackpal agency upsellFact plus opportunity
Claim/fact-check ingestionQuery Google Fact Check, normalise reviews, dedupe, cluster claims, track publishers, and record safe fetch attempts.Bollocks RadarLocal implementation / documented patternMediumProduct-local first; source adapter pattern laterBollocks Radar, public trust widgets, SiteIntel trust signalsFact

Provider API sync and external integrations

CapabilitySummarySource projectsMaturityReuse potentialLikely extraction shapeProducts it could helpEvidence status
Read-only business metric syncConnect provider accounts, fetch top-level metrics, write metric snapshots, and record sync runs.TrackpalLocal implementationMedium to highConnector interface and sync-run patternTrackpal, future small-business SaaS productsFact
Provider rollout registryTrack provider setup status, required config, required secret names, scopes, approval blockers, templates, and metrics.Trackpal, Meridian Hub, SiteIntelRepeated patternHighShared contract and admin componentsTrackpal, SiteIntel, DataStitcher, HubFact plus inference
Secret presence/status surfacesShow secret/config status as present/missing/unknown/not expected/stale without exposing values.Trackpal, Meridian Hub, Meridian deployment/runtime docsDocumented patternHighWorks admin component plus CLI/runtime helperEvery protected admin surfaceFact
BigQuery metadata syncRead project, dataset, table, schema, and change metadata from BigQuery APIs or INFORMATION_SCHEMA with bounds.BigQuery Atlas, SiteIntel BigQuery worker, Meridian works-bigquery-catalogueRepeated patternHighPackage or CLI adapter with clear safety docsBigQuery Atlas, DataStitcher, SiteIntel, Compass generationFact

Dashboard, admin UI, and public reporting

CapabilitySummarySource projectsMaturityReuse potentialLikely extraction shapeProducts it could helpEvidence status
Template-driven scorecardsRender dashboards from template definitions, metric definitions, snapshots, alerts, freshness, and setup prompts.TrackpalLocal implementation / documented patternHighPattern first, then React view-model packageTrackpal, DataStitcher, admin scoreboards, future SaaS appsFact plus opportunity
Evidence-backed story reportsPresent deterministic findings in plain English with source evidence and optional technical proof.SiteIntel, Bollocks RadarRepeated product patternHighGuide plus finding/evidence contractsSiteIntel, Bollocks Radar, public audit toolsFact plus inference
Protected admin operations shellProvide dense operator surfaces for invites, sources, sync runs, provider health, manual actions, and feature flags.Trackpal, SiteIntel, DataStitcher, BigQuery Atlas, Meridian HubRepeated patternHighBlueprint first, then component libraryMost current productsFact
Generated docs cockpitsGenerate status, operations, deployable-unit, runtime, and live-data docs from checked-in metadata.Product repos, @meridian-stack/works-docs, Product CompassExtracted component / governed package for docsHighExisting package plus product metadata conventionsEvery product repo and Product CompassFact

Data modelling, registry, and governance

CapabilitySummarySource projectsMaturityReuse potentialLikely extraction shapeProducts it could helpEvidence status
Semantic metric registryDefine metrics, dimensions, aliases, caveats, external mappings, sensitivity, and publication rules.DataStitcher, TrackpalRepeated patternHighRegistry schema and docs-generation guideDataStitcher, Trackpal, public data toolsFact plus inference
Domain registry and Public Suffix List syncStore canonical domain parts, ICANN/full PSL registrable domains, hierarchy edges, and source sightings.SiteIntelLocal implementation / prototype docsHighPackage or BigQuery schema template after proofSiteIntel, Trackpal, domain tools, public audit toolsFact plus opportunity
Data product separationSeparate platform capabilities, data product config, and public app code.DataStitcherDocumented patternMedium to highArchitecture guide and starter blueprintDataStitcher, future public-data productsFact
Privacy and publication policyEncode whether raw/name-bearing data can be stored or published, including region constraints.DataStitcherDocumented patternMediumGuide and schema fieldsDataStitcher, public audit tools, Bollocks RadarFact

AI-agent support and developer intelligence

CapabilitySummarySource projectsMaturityReuse potentialLikely extraction shapeProducts it could helpEvidence status
AI context pack generationGenerate compact Markdown/JSON context from project facts and synthesis docs.Product Compass, product docs reposLocal implementationHighImproved script plus metadata conventionsProduct Compass, future AI handoff flowsFact
Codex session analysisParse local Codex JSONL, derive costs/timing/tool timelines, render reports and dashboards, export redacted metrics.Codex Sessions AnalyserLocal implementationHighLocal developer tool, possibly redacted Hub feedMeridian, Compass, AI-agent workflow improvementFact
Repository analysis reportsRun context audit in GitHub Actions, post full/compact reports to Hub, store full reports in R2 and compact summaries in D1.Meridian HubLocal implementation / documented patternHighHub feature plus possible Compass inputMeridian Hub, Product Compass, AI onboardingFact

First extraction priorities

  1. Ingestion contracts: source, run, step, artefact, validation, retry, and status shapes.
  2. Admin status components: provider status, secret-status, sync/run summaries, recent failures, manual retry controls.
  3. Source/provider registry contract: shared language for enabled state, cadence, credential requirements, scopes, and supported metrics.
  4. Domain/Public Suffix List package or blueprint: extract only after SiteIntel and a second consumer need it.
  5. Template dashboard renderer: keep Trackpal-local until another product renders from the same config shape.

What should remain local for now

  • Bollocks Radar verdict mapping, claim clustering, and exact Google Fact Check ingestion semantics.
  • Trackpal live provider connector details and OAuth flows.
  • DataStitcher business-specific metrics and ABN publication rules.
  • SiteIntel findings copy, competitor interpretation, and paid-source selection.
  • BigQuery Atlas scoring heuristics until candidate discovery and validation prove stable under real runs.

Sources consulted

  • projects/site-intel/docs/architecture/task-orchestration.mdx
  • projects/site-intel/docs/reference/technical/bigquery-domain-registry-and-bulk-ingest.mdx
  • projects/data-stitcher/docs/operations/reusable-monitored-ingestion-pattern.md
  • projects/data-stitcher/docs/reference/registry/overview.mdx
  • projects/bq-atlas/docs/architecture/shared-code-extraction-plan.md
  • projects/bq-atlas/services/candidate-crawler/README.md
  • projects/bq-atlas/services/public-cataloguer/README.md
  • projects/bollocks-radar/docs/reference/technical/google-fact-check-ingestion.mdx
  • projects/trackpal/docs/product/07-template-system.md
  • projects/trackpal/docs/product/12-integrations-strategy.md
  • projects/codex-sessions-analyser/README.md
  • docs/architecture/meridian-works-shared-platform.md