Skip to content

Release v19.8.5

Choose a tag to compare

View all tags
@github-actions github-actions released this 17 Jun 00:20
· 38 commits to main since this release
v19.8.5
618cb50

What's Changed

  • refactor(agent-sync): raw-string block, drop dead Surface field, derive docCount (618cb50)
  • docs(agent-sync): correct stale projector docstring + design-doc nits (ea31d1f)
  • docs(agent-guidance): record final split + plugin retirement (4faf020)
  • chore(agent-guidance): remove GitHub plugin marketplace packaging (07e0566)
  • docs(spiderly-upgrade): run agent-sync to refresh projected guidance (ce96d7a)
  • refactor(agent-surface): testing skills -> docs; add-entity command -> skill (a0a438d)
  • feat(agent-sync): static docs pointer + junction from split skills dir (0b3d2b4)
  • refactor(agent-bundle): split bundle into docs/ + skills/ by surface (5152880)
  • docs(plan): agent-guidance distribution finalize (phase 3 + split) (a943b66)
  • ci: add CODEOWNERS for owner-reserved paths (85581a4)
  • ci: migrate claude-review to reusable ci-workflows/claude-gate.yml@v1 (67716ae)
  • docs(entity-design): clarify validation attributes on ReadonlyObject entities (047b565)
  • ci(release): note why dry-run-summary must declare the publish jobs in needs (c17d39a)
  • ci: add Claude Code review + @claude workflows (76dcb3f)
  • ci(agent-evals): add out-of-band eval workflow + harness self-test gate (0d95f80)
  • feat(agent-evals): make add-validator a real, discriminating task (369a006)
  • Fix Enter key not submitting the login verification code (37b8b33)
  • refactor(agent-evals): thread tracksByName + cleaner report sort (49b6843)
  • fix(agent-evals): address final-review issues (loud CLI validation, grep removal) (10e1752)
  • feat(agent-evals): first real task add-validator (dotnet-build verifier, out-of-band) (39b7104)
  • feat(agent-evals): CLI entrypoint + real Claude headless adapter (93efd3b)
  • feat(agent-evals): markdown report renderer with failure digest (0bcf64f)
  • feat(agent-evals): scorer + orchestrator + oracle/no-op self-test gate (9bea455)
  • feat(agent-evals): no-op and oracle fake agents + trivial oracle patch (494caf4)
  • feat(agent-evals): verifier runner + trivial task verifier (d849811)
  • feat(agent-evals): agnostic track provisioning (fixture + AGENTS.md index) (c7f03b1)
  • feat(agent-evals): task loader + trivial-marker self-test task (0189497)
  • chore(agent-evals): fix node --test invocation for Node 24 (glob, not dir) (66c07e2)
  • feat(agent-evals): scaffold harness paths, fs/exec helpers, trivial fixture (4b43fc5)
  • docs: add agent eval harness Phase 1 implementation plan (dc59a0b)
  • feat(agent-guidance): ship version-matched docs+skills in the npm package via spiderly agent-sync (65a8233)
  • docs: add agent eval harness design spec (acfde5e)
  • Pass ActionClickEvent payload to data-table Action.onClick (5ff3125)
  • docs(skills): surface optimistic concurrency where save-pipeline reasoning happens (c1b8d24)
  • docs(skills): document global spinner + global error handling — stop per-call boilerplate (49a4b2c)
  • docs(custom-endpoints): validation attributes are for input DTOs, not readonly ones (928b0a3)
  • feat(deployment-skill): trigger on deploy/redeploy/incident intents, not just setup (2f2b9aa)
  • fix(deployment-skill): make CI sequence first-cutover-safe and idempotent (d8654a3)
  • fix(deployment-skill): mount postgres volume at /var/lib/postgresql for 18+ (ce724db)
  • feat(report-gap): fire on gaps in Spiderly's own skills/plugins/docs (cfe07a7)
  • docs(report-gap): agent-reported label already exists — stop instructing creation (af1615c)

Packages Published

NuGet Packages

npm Package

Full Changelog: v19.8.4...v19.8.5

Contributors

claude
Assets 7
Loading