Proposals Directory

Type: kb/types/index.md

← Parent

Subdirectories

Files

  • A reader-facing banner for user verification (design-proposal) - Proposal: surface committed user verification at the point of reading through a standardized Markdown banner without turning absence into a review or truth verdict
  • Ablation baselines for the declared objective (design-proposal) - Proposal: measure the declared objective — useful and warranted work per unit of human judgment — by running matched repository tasks under ablated conditions (no curated theory, no review, episodes only, stronger bare model)
  • Automated note refinement as a search over a fixed source bundle (design-proposal) - Proposal: build the automated note-refinement loop on a source bundle that emits a set of notes, not a single identity-stable note — reframing non-convergence (split, drift, kill) as search outcomes
  • Backlink surfacing (design-proposal) - Proposal: surface inbound links through build-time web rendering for humans and on-demand inversion for agents while allowing useful reciprocal links, not mandatory mirroring
  • Bidirectional codification as a comparative test (design-proposal) - Proposal: promote bidirectional codification from design guidance to a comparative conjecture, tested by evolving the same task stream under natural-language-only, symbolic-only, one-way promotion, and bidirectional codify-and-relax regimes
  • Calibrating semantic gates against labelled fixtures (design-proposal) - Proposal: require non-leaking known-case regression before a semantic gate can advance, and reserve live detection-rate claims for separately sampled field calibration
  • Channel-compiled instruction artifacts (design-proposal) - Proposal: resolve execution channel at install time so an agent reads one literal procedure, extending build-time generation from paths to shell and command form
  • Checked inline blocks for shared instruction text (design-proposal) - Proposal: reuse natural-language authoring mechanics in specialized writer prompts through literal inlining backed by deterministic source-to-copy checks
  • Collection-as-artifact freshness (design-proposal) - Proposal: register collection-maintenance targets with collection-text inputs for casebook-wide staleness without per-file dependency edges
  • Factored dependency pairs for review freshness (design-proposal) - Proposal: admit new review dependencies as factored (note, dependency) pairs with the dependency as the gate — source-as-gate and a cohort-scoped ack remain; COLLECTION.md-as-gate adopted (ADR 041)
  • Gate learning from accepted edits (design-proposal) - Proposal: learn review gates from accepted note edits, with candidate mining, usefulness tracking, lifecycle promotion, and budgeted loading. Atomic gates shipped; learning has not
  • Generalized validation invalidation and imperative extension (design-proposal) - Proposal: generalize validator invalidation or imperative type extension only after explicit selectors and local mark cases prove reusable machinery
  • Model partition registry (design-proposal) - Proposal: introduce a model partition registry for review validation, aliases, and runner defaults without making the registry the review identity
  • Periodic connect-report mining (design-proposal) - Proposal: mine connect reports on a recurring cadence as a bulk-operation pipeline, automating the noticing-to-candidate triage step connect and kb/log.md leave unattended
  • Review configuration object (design-proposal) - Proposal: give commonplace.review a ReviewConfig for review scan roots, gate and artifact locations, and database path with project overrides, without introducing a global ProjectConfig
  • Revise the behavioral-authority decomposition (design-proposal) - Proposal: restructure the behavioral-authority sub-divisions — pipeline paths, force families, an effective/warranted split — with a literature survey of existing authority and provenance decompositions before any carve is chosen
  • Structured-output codec for the review protocol (design-proposal) - Proposal: make review output a pluggable codec, with sentinel markdown today and schema-validated structured output when harnesses expose it
  • Tag scope is declared where membership claims are made (design-proposal) - Proposal: attach tag scope to membership claims (marks, sweep recipes) rather than to the tag string — options: formalized collection scope, repo scope, or a declared per-README dual scope
  • Trajectory-aware evaluation of transforming agent workflows (design-proposal) - Proposal: compare blinded output-only and trajectory-aware judges on transforming agent workflows before adding trajectory evaluation to closure
  • Where subtree-scoped write-time contracts live (design-proposal) - Proposal: where a write-time contract binding less than a whole collection lives — COLLECTION.md fold, local type, nested contracts, or a new area surface; the instance's core since typed
  • Write-time vocabulary collision controls (design-proposal) - Proposal: mechanical controls for the one-term-one-sense invariant — reserved-term registry, slot-escape lint, coinage collision screen, naming-review gate, and clausal-binding link check