Skip to content

feat(workspace-probe): @metaharness/workspace-probe β€” eval + Darwin-Mode bridge for workspace-lens (ADR-239)#131

Merged
ruvnet merged 1 commit into
mainfrom
feat/workspace-probe
Jul 7, 2026
Merged

feat(workspace-probe): @metaharness/workspace-probe β€” eval + Darwin-Mode bridge for workspace-lens (ADR-239)#131
ruvnet merged 1 commit into
mainfrom
feat/workspace-probe

Conversation

@ruvnet

@ruvnet ruvnet commented Jul 7, 2026

Copy link
Copy Markdown
Owner

What / why

A thin, pure companion package to @metaharness/workspace-lens (ADR-238, #129) β€” the application layer that turns Jacobian-Lens interpretability receipts into evaluation signal:

  • workspaceProbeScore(receipts, {driftThreshold}) β†’ { n, score = cleanFraction, meanDrift, flagRate, criticalRate } β€” a flywheel-consumable workspace_probe surface: the fraction of decisions whose workspace was clean (no critical trigger AND drift below threshold). Empty set β†’ 0 (nothing witnessed β†’ no credit).
  • gradeMutationByWorkspace(baseline, mutant) β†’ { keep, reasons, baseline, mutant } β€” Darwin-Mode mutation evidence: a veto that rejects a mutation which raises the critical-trigger rate, destabilizes the workspace (mean drift up beyond tolerance), or drops the clean fraction β€” even if final answers improved ("final answer up, workspace grip down" = structurally brittle). It pairs with the gold/answer gate (keep only if both pass) and never weakens it; meetsPromotionRule is untouched.

Depends only on the workspace-lens types; deterministic; $0.

Validation

  • Logic developed + validated in isolation against the published @metaharness/workspace-lens@0.1.0, then re-homed with a local workspace dep.
  • 9 synthetic tests (probe scoring, empty-set, custom threshold, all four mutation-veto paths, helpers); tsc clean against the local workspace-lens; healthcheck 8/8.
  • Wired into build-ordered phase-2 (builds after workspace-lens in phase-1) + the version-coherence allowlist; ADR-239.
  • The lockfile also gains the @metaharness/workspace-lens workspace entry that feat(workspace-lens): @metaharness/workspace-lens β€” Jacobian-Lens interpretability primitive (ADR-238)Β #129 didn't add.

πŸ€– Generated with claude-flow

…ode bridge for workspace-lens (ADR-239)

A thin, pure companion to @metaharness/workspace-lens (ADR-238) that turns Jacobian-Lens interpretability
receipts into evaluation signal:

- workspaceProbeScore(receipts, {driftThreshold}) β†’ { n, score=cleanFraction, meanDrift, flagRate,
  criticalRate } β€” a flywheel-consumable `workspace_probe` surface: the fraction of decisions whose
  workspace was clean (no critical trigger AND drift below threshold). Empty set β†’ 0 (nothing witnessed
  β†’ no credit).
- gradeMutationByWorkspace(baseline, mutant, opts) β†’ { keep, reasons, baseline, mutant } β€” Darwin-Mode
  mutation evidence: a VETO that rejects a mutation which raises the critical-trigger rate, destabilizes
  the workspace (mean drift up beyond tolerance), or drops the clean fraction β€” even if final answers
  improved ("final answer up, workspace grip down" = structurally brittle). Pairs with the gold/answer
  gate (keep only if BOTH pass); never weakens it. meetsPromotionRule untouched.

Depends only on the workspace-lens types; deterministic; $0. Logic was developed + validated in
isolation against the published @metaharness/workspace-lens@0.1.0, then re-homed with a local workspace
dep. 9 synthetic tests; tsc clean; healthcheck 8/8. Wired into build-ordered phase-2 (after workspace-lens)
+ the version-coherence allowlist; ADR-239. (The lockfile also gains the workspace-lens entry that #129
did not add.)

Co-Authored-By: claude-flow <ruv@ruv.net>
@ruvnet ruvnet merged commit 6e7d741 into main Jul 7, 2026
25 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant