Skip to main content

Source Ledger and Editorial Standard

Last technically verified: 2026-08-09

Stability Labels

LabelMeaningEditorial treatment
DurableExpected to outlive individual productsTeach in the main lifecycle
EvolvingStandard or practice is still changingCite the applicable specification/version
SnapshotCommands, versions, support, pricing, modelsKeep in dated appendices

Source Record

Every time-sensitive claim records: claim, source URL, source type, verification date, applicable version, stability, and affected chapter.

TopicPrimary sourceTypeVerifiedStabilityUsed in
Spec Kit lifecyclehttps://github.com/github/spec-kit/blob/main/docs/reference/agentic-sdd.mdOfficial docs2026-08-09EvolvingPart VI
Spec Kit workflowshttps://github.com/github/spec-kit/blob/main/docs/reference/workflows.mdOfficial docs2026-08-09EvolvingParts VI, XII
OpenSpec OPSXhttps://openspec.dev/docs/opsxOfficial docs2026-08-09EvolvingPart VI
OpenSpec CLI/schemashttps://openspec.dev/docs/reference/cliOfficial docs2026-08-09EvolvingPart VI
BMAD workflowhttps://github.com/bmad-code-org/BMAD-METHOD/blob/main/docs/reference/workflow-map.mdOfficial repository2026-08-09EvolvingPart VI
AGENTS.mdhttps://agents.md/Standard site2026-08-09DurablePart VII
MCPhttps://modelcontextprotocol.io/specification/2025-11-25Protocol specification2026-08-09EvolvingPart VIII
A2Ahttps://a2a-protocol.org/latest/specification/Protocol specification2026-08-09EvolvingPart VIII
Agent Skillshttps://docs.github.com/en/copilot/concepts/agents/about-agent-skillsOfficial docs2026-08-09EvolvingPart VII
OWASP agent securityhttps://owasp.org/www-project-top-10-for-large-language-model-applications/Security standard2026-08-09EvolvingPart IX
NIST AI RMFhttps://www.nist.gov/itl/ai-risk-management-frameworkGovernment standard2026-08-09DurableParts IX, XIV
OpenTelemetryhttps://opentelemetry.io/docs/specs/Specification2026-08-09EvolvingPart XIII
SLSAhttps://slsa.dev/spec/Supply-chain standard2026-08-09EvolvingPart XII

Citation Rules

  • Prefer specifications, official documentation, repositories, standards bodies, and peer-reviewed research.
  • Use vendor blogs only for the vendor's own product behavior.
  • Never turn marketing claims into general industry conclusions.
  • Mark previews and experimental protocol features explicitly.
  • Re-run documented commands before updating a snapshot.
  • Treat Firecrawl and Tavily output as discovery, not final authority.