Release: dev -> main - #249
Merged
Merged
Conversation
Update the Pi review CLI so source prompts treat pasted text as refine feedback, winner prompts use s1-s5 or rel for scale adjustment, and derive remains a numbered option. Also make quit stop active round coordinators, mark those rounds as stopped, and let later resume continue from completed work without restarting the queue.
Add a broad long-horizon research program for evolving a local agent harness around a base model, including retrieval policy, symbolic behavioral threads, tool use, memory, and evaluation design. This frames the MSI lane as an evolutionary outer loop over agent packages rather than a direct frontier-model training effort.
Move the training-prompts program and prompt catalog under dev-research/training-prompts. Update the Pi review runtime, workflow helpers, tests, and docs to use the new paths.
Add deterministic skill-link extraction and encoding validation tools to the main CLI. Clean up skill discovery and validation paths. Align the behavioral-core and hypergraph-memory skills with current Plaited runtime behavior. Add the default-hypergraph and behavioral-factories research programs and seed graph artifacts. Add the Pi fanout research harness for worktree-backed runs.
Rename mss-vocabulary to mss across scripts, programs, and skill references. Audit the MSS skill and distilled references. Treat boundary and scale as the strongest invariants. Reframe contentType, structure, and mechanics as alignment-driven in agent-mediated systems. Add a deterministic mss-doc-chunks script and test for chunking the large source docs with provenance. Update AGENTS.md to assess area of effect before choosing targeted or full validation.
Add scripts/mss-source-compare.ts as a deterministic MSS source comparison surface with optional embeddings, m2.7 pair analysis, and glm concept proposals. Wire the new runner into package.json and add focused tests for signatures, pair selection, prompts, and disposition heuristics.
Default the optional embedding pass in mss-source-compare to nvidia/llama-nemotron-embed-vl-1b-v2:free.
Clarify the default-hypergraph and behavioral-factories programs so they stay conceptual and leave concrete harness, skill, and search-provider wiring to the fanout runner. Update research-pi-fanout to use a total attempt budget of 15, an initial concurrent fanout of 3, and expose the You.com search skill to default-hypergraph attempts.
Keep the root default-hypergraph and behavioral-factories scripts as stable validator surfaces while moving lane-owned augmentation to program-local script directories under dev-research. Narrow fanout writable roots to each lane tree and retain retry guidance when attempts edit outside the allowed program surface.
Resolve varlock from the repo root for fanout attempts and keep Pi session state inside each attempt directory. Also teach the default-hypergraph and behavioral-factories programs to use validators, typecheck, targeted tests, and Biome as part of their deterministic assessment loop.
Add internal reference validation for the default-hypergraph seed, include untracked files in fanout attempt diffs, and tighten the lane instructions around stable support surfaces. Promote the strongest default-hypergraph seed additions from the recent fanout run, including runtime, invariant, distillation, factory, and boundary safety artifacts plus targeted MSS, modnet, and agent-policy coverage improvements.
Shrink the default-hypergraph lane back to a smaller seed-only contract. Rename seed document ids away from skill://. Clean up validator terminology and checks to match the seed-document model. Trim ontology overgrowth and remove stale side artifacts. Update MSS scale descriptions and switch repo hooks to bunx.
Integrate autoresearch lane evaluation with improve graders and meta-verifiers, rename the shared repo-improvement eval surface to workspace-improvement, and add MSS seed/corpus judge and verifier implementations. Document external retrieval for the MSS lanes, provision the You.com skill as a fallback capability during autoresearch runs, and update operator guidance in AGENTS.md.
Update the README to reflect the current autoresearch and MSS lane surfaces, and move the obsolete root-level native model proof plan into docs/archive so it remains preserved without looking active.
Resolve the active workspace root from git instead of assuming the current shell cwd, so attempt worktrees write lane outputs under their own dev-research surfaces. Keep varlock and skill resolution anchored to the canonical repo root while passing the attempt worktree as PLAITED_WORKSPACE_ROOT for lane scripts and validators. Update the MSS lane tests to cover workspace-root overrides and runner run-dir construction.
Add Pi worktree path guards. Add workspace-rooted judge file access. Commit validated attempts before judging. Add root-level promotion selection and stronger targeted tests.
- index.jsonld, modnet-core.jsonld, mss-composition.jsonld - mss-dynamics.jsonld, mss-fields.jsonld, mss-invariants.jsonld - mss-runtime.jsonld, mss-values.jsonld These JSON-LD seed anchors are required by the mss-corpus lane.
- chunks.jsonl: 141 source chunks from Structural-IA.md, Modnet.md, MODNET-IMPLEMENTATION.md with provenance links (sourcePath, heading, headingPath, kind) - source-compare.json: 30 deterministic pairs linking skill docs to chunks with sharedTerms, scores, and preserve/reinterpret/archive disposition - encoded/manifest.json: manifest linking corpus to seed anchors Corpus artifacts align to mss-seed anchors (five fields, invariants, scale-structure maps) and provide graph-ready inputs for downstream behavioral-factories lane.
Wrap MSS autoresearch entrypoints with varlock. Keep judges and promotion selection on the same secret env as Pi. Remove committed seed and corpus outputs before the sandbox rework.
Block nested autoresearch and canonical repo-root bash targets. Keep worktree-local writes allowed. Pass the canonical repo root into the Pi guard and expand targeted tests.
Add behavioral-seed and behavioral-corpus lane programs and script entrypoints, wire behavioral-factories into lane evaluation, tighten workspace-improvement judge context, and move constitution governance framing into skill references. Also remove the stale Modnet human CLI spec and collapse its live workflow guidance into the training-prompts program.
Remove docs/MODNET-IMPLEMENTATION.md and move the remaining modnet topology, A2A, and access-control guidance into the modnet-node skill and its references. Update the MSS research lanes and supporting docs to point at those skill surfaces instead of the removed design doc. Also keep the in-flight semantic lane validation changes and modnet-modules asset cleanup in the same commit so the repo state remains internally consistent.
Add research program surfaces for agent-loop, node-auth, project-isolation, and proactive-node so the exploratory architecture can evolve separately from the stable skill contracts. Trim the skills toward durable guidance and move the still experimental policy questions into dev-research programs.
Align the behavioral seed and corpus lanes with the newer MSS-style evaluation flow by adding graders, verifiers, and updated validation coverage. Reposition trial-runner and trial-adapters as public improve surfaces, remove the stale training-pipeline skill, and fold the relevant direction into evolutionary-agent and training docs. Soften constitution reference prose to treat factory patterns, generated bThread structure, and MAC rules as current baselines and starting points rather than final doctrine.
Add a manual workflow that validates the release-readiness issue. It opens or updates the dev-to-main release PR without merging, publishing, or mutating branches. Update the repo-local development skill to document this lane.
Exclude exact release-readiness and open-release-pr GitHub Actions check-runs from release-readiness aggregation so expected operator workflow failures do not block release readiness.
github-actions
Bot
requested review from
EdwardIrby and
alisonailea
as code owners
April 15, 2026 01:37
![github-advanced-security[bot]](https://avatars.githubusercontent.com/in/57789?s=60&v=4){kind=small}
|
|
||
| - name: Setup Bun | ||
| if: ${{ steps.gate.outputs.should_run == 'true' }} | ||
| uses: oven-sh/setup-bun@v2 |
Pin oven-sh/setup-bun to an immutable commit in cline-pr-review. This resolves the CodeQL unpinned-action finding on the release PR.
EdwardIrby
approved these changes
Apr 15, 2026
EdwardIrby
requested review from
a team and
EdwardIrby
and removed request for
EdwardIrby and
alisonailea
April 15, 2026 01:57
EdwardIrby
approved these changes
Apr 15, 2026
Member
|
Refreshing PR event processing to clear stale Code Quality pending-language gate. |
Treat same-repo PRs as trusted and only run external intake closure/comment logic for external fork PRs.
EdwardIrby
self-requested a review
April 15, 2026 02:14
EdwardIrby
previously approved these changes
Apr 15, 2026
EdwardIrby
approved these changes
Apr 15, 2026
7 tasks
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Context
devtomain.Summary
Updated to current
devSHA6ad32fde211554b092867fa9ce4e7ba2eb577247.Includes fix
6ad32fde: external-pr-intake now scopes auto-close/comment to external fork PRs only.Opened by the manual
Open Release PRworkflow after release-readiness validation.Full auto-generated readiness packet is included below.
Changed Files
devintomain.Validation
Known Failures / Drift
Review Notes / Residual Risks
dev -> mainrelease scope.main -> devmerge-commit sync.Agent Workflow Checklist
main -> devsync requirement retainedRelease Packet (Auto-generated)
Context
This is the human-gated release PR from
devtomain.Release Readiness
devmainb7921c330304d5a014ee0bdad9349b7d199aeaf5Readiness Packet
Release Notes Draft
Validation Summary
Deterministic Security Summary
release-readiness-secretavailableavailableavailableavailable{ "medium": 3 }{}0extendedDeterministic Check Summary
b7921c330304d5a014ee0bdad9349b7d199aeaf5available451801530001Included PRs (Deterministic Match)
Changed Surfaces
P0/P1 Checklist
Human Decision Checklist
Post-Merge Requirement
After this PR is squash-merged into
main, syncmainback intodevwith a merge commit.Never reset/rebase/force-push
dev.Human Approval
This PR must be reviewed and merged by a maintainer. The workflow does not auto-merge or
publish.