Skip to content

feat(scient-agent): drive Scient Agent 1.x, built on Oh My Pi 19 - #1034

Merged
yaacovcorcos merged 3 commits into
mainfrom
claude/scient-agent-omp19-20261120
Nov 20, 2026
Merged

yaacovcorcos merged 3 commits into
mainfrom
claude/scient-agent-omp19-20261120

Conversation

@yaacovcorcos

Copy link
Copy Markdown
Contributor

What Changed

Scient Desktop drives Scient Agent 1.x, the agent aligned with Oh My Pi 19.0.1 (ScientFactory/scient-agent#77). Three commits:

  • Admission. The Scient Agent target accepts product majors 0 and 1, and judges the agent's upstream release by its own range: Oh My Pi 18.2.8 or newer on major 18 or 19. Managed installation accepts Scient Agent below 2.0.0. The stock Oh My Pi provider keeps major 18; isSupportedOmpRuntimeVersion is no longer exported, since only OmpTarget uses it.
  • Resume. A Scient Agent resume cursor written on Oh My Pi 18 resumes on 19 (OmpTarget.resumableUpstreamMajors), never the reverse. Stock Oh My Pi keeps the same-major rule.
  • Compatibility banner. The model manifest's Scient Agent policy is split by Scient release: up to 0.6.31 it keeps <1.0.0; from 0.6.32 it supports >=0.1.0 <2.0.0. updatedAt moves forward so a cached manifest from 0.6.31 does not shadow the new bundle.

Why

Every shipped Desktop refuses a Scient Agent built on Oh My Pi 19: its target required the stock provider's supported major, 18, for the agent's upstream release. Desktop's live suites against the 19.0.1 candidate failed every session open with "This is not a Scient Agent release this version of Scient supports".

The 19-based agent is released as Scient Agent 1.0.0, so Desktops that cannot run it never install it: their managed-runtime ceiling stops below 1.0.0, and their target accepts only 0.x. They keep 0.1.13. This was weighed against bumping the catalog contract revision (a Codex debate): the revision bump would have left a fresh install of this Desktop without an installable agent until the new agent was promoted. The agent's release check now refuses any other version for this build (scient-agent#77).

Without the resume change, every thread last run on 0.1.13 would start a fresh native session from a recovery summary after the update. Oh My Pi 19 reads an 18 session: Scient Agent's capability check resumes and continues a session written by 0.1.13 on the 19.0.1 build.

Without the manifest change, installed releases (which read the manifest from main) would mark a working 1.0.0 as unsupported, or this release would.

Release order: this merges and ships in a Scient release before Scient Agent 1.0.0 is released and promoted in the managed-runtime catalog. If a Scient release other than 0.6.32 is cut from main first without this change, the manifest's 0.6.32 bound moves to that release's successor.

Known limit, unchanged by any gate: once the catalog promotes 1.0.0, a fresh install of a Scient release from before this change, with no installed or cached Scient Agent, has no agent to install until it updates itself.

Documentation Impact

Documentation impact: Updated — docs/internals/providers.md (the identity range and resume across majors).

Platforms

Platform Expected behavior Checks run Unproven
macOS Same on every platform: version checks and JSON policy, no platform code. Focused unit tests, typecheck, knip, lint, format; live suites against the 19.0.1 candidate and v0.1.13 on macOS arm64. Intel.
Windows Same. CI. Live suites.
Linux Same. CI. Live suites.

Verification

  • apps/server: src/provider/scient, src/provider/omp, src/scient/providerLifecycle: 857 passed. packages/scient-provider-runtime: 403 passed. scripts/lib/managed-runtime-catalog.test.ts: 75 passed. providerCompatibility.test.ts and ModelManifest*: 38 passed.
  • pnpm run typecheck, pnpm run knip:check, vp lint, vp fmt --check, git diff --check, scripts/scient-seam-check.mjs, scripts/verify-upstream-provenance.mjs: pass.
  • Live suites (OMP_QUALIFY_TARGET=scient, 12 files) on this head: the 19.0.1 candidate and v0.1.13 fail the same 13 of 34 cases and no other (this machine's Ollama lacks the custom-model fixture's model, no stock omp runs beside it, and the Scient extension fixture's tool does not load for either). On main the candidate failed 25, every session refused.
  • Not run: the full suite, pnpm run build, desktop smoke; CI runs them.

Review

Review: Codex high and a live test — 3 findings on the first head: 2 fixed (resume across majors, the manifest's banner), 1 rejected (a fresh install of an old release after promotion: any gate has it, and the app updates itself); recheck found the manifest's updatedAt left unchanged, fixed; recheck otherwise clean.

Checklist

  • This PR is small and focused
  • I explained what changed and why
  • Focused tests and typecheck pass on this head
  • UI: no UI change
  • Every link a reader can follow that this PR adds was fetched and works

Model/harness: Claude Opus 5.5 (Claude Code), reviews by Codex gpt-6.2-sol.

🤖 Generated with Claude Code

yaacovcorcos and others added 3 commits November 20, 2026 02:37
Scient Agent has been aligned with Oh My Pi 19.0.1. Every shipped Desktop
refuses that build: the Scient Agent target required the stock Oh My Pi
provider's supported major, 18, for the agent's upstream release.

Scient Agent is qualified release by release, so its upstream range is now
its own: Oh My Pi 18.2.8 or newer on major 18 or 19. The 19-based agent ships
as Scient Agent 1.0.0, a new product major, because Desktops that cannot run
it must never install it: their managed-runtime policy stops below 1.0.0 and
their target accepts only 0.x. This Desktop accepts 0.x and 1.x and installs
below 2.0.0. The stock Oh My Pi provider stays on 18.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A resume cursor records the Oh My Pi release that wrote it, and a cursor from
another major was refused: after updating to Scient Agent 1.x, built on Oh My
Pi 19, every thread last run on 0.x would start a fresh native session from a
recovery summary. Oh My Pi 19 resumes and continues an 18 session (Scient
Agent's capability check runs a 0.1.13 session on the new build), so Scient
Agent's target lists 18 then 19 as majors that resume forward. A cursor still
never resumes on an older major, and stock Oh My Pi keeps the same-major rule.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Installed Scient releases read the model manifest's compatibility policies from
main, so its Scient Agent entry keeps the 0.x range for releases up to 0.6.31,
which cannot run 1.x, and a new entry from 0.6.32 on supports Scient Agent
below 2.0.0. Without it the release that drives 1.0.0 would show it as an
unsupported version.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@github-actions github-actions Bot added vouch:trusted Author is trusted by repo permissions or the VOUCHED list. size:M labels Nov 20, 2026
@yaacovcorcos
yaacovcorcos added this pull request to the merge queue Nov 20, 2026
@github-actions

Copy link
Copy Markdown
Contributor

Thread transfer impact

✅ Thread transfer remains within every enforced ceiling.

Provider Metric Main baseline This PR Impact PR ceiling
Codex Total thread wire 5.5 KiB 5.5 KiB 0 B (0.0%) 6.8 KiB ✅
Codex Thread snapshot wire 4.2 KiB 4.2 KiB 0 B (0.0%) 4.9 KiB ✅
Codex Live turn WebSocket wire 1.3 KiB 1.3 KiB 0 B (0.0%) 2.0 KiB ✅
Codex Live turn WebSocket decoded 22.3 KiB 22.3 KiB 0 B (0.0%) 24.4 KiB ✅
Codex Live turn messages 1 1 0 (0.0%) 8 ✅
Claude Total thread wire 5.5 KiB 5.5 KiB 0 B (0.0%) 6.8 KiB ✅
Claude Thread snapshot wire 4.2 KiB 4.2 KiB 0 B (0.0%) 4.9 KiB ✅
Claude Live turn WebSocket wire 1.3 KiB 1.3 KiB 0 B (0.0%) 2.0 KiB ✅
Claude Live turn WebSocket decoded 22.7 KiB 22.7 KiB 0 B (0.0%) 24.4 KiB ✅
Claude Live turn messages 1 1 0 (0.0%) 8 ✅

Other scenarios

All 64 scenario measurements match main.

All 64 scenario measurements
Provider Scenario Metric Main baseline This PR Impact PR ceiling
Codex short turn resume WebSocket decoded 12.0 KiB 12.0 KiB 0 B (0.0%) 12.7 KiB ✅
Codex short turn resume WebSocket messages 2 2 0 (0.0%) 2 ✅
Codex long turn running, cold open HTTP decoded 281.7 KiB 281.7 KiB 0 B (0.0%) 302.7 KiB ✅
Codex long turn running, cold open rows 75 75 0 (0.0%) 75 ✅
Codex long turn running, history page HTTP decoded 92.4 KiB 92.4 KiB 0 B (0.0%) 107.4 KiB ✅
Codex long turn running, history page rows 75 75 0 (0.0%) 75 ✅
Codex long turn finished, cold open HTTP decoded 381.5 KiB 381.5 KiB 0 B (0.0%) 410.2 KiB ✅
Codex long turn finished, cold open rows 151 151 0 (0.0%) 151 ✅
Codex long turn finished, socket cold open WebSocket decoded 381.6 KiB 381.6 KiB 0 B (0.0%) 410.2 KiB ✅
Codex long turn finished, socket cold open WebSocket messages 2 2 0 (0.0%) 2 ✅
Codex long turn finished, socket cold open rows 151 151 0 (0.0%) 151 ✅
Codex long turn finished, work page HTTP decoded 186.2 KiB 186.2 KiB 0 B (0.0%) 200.2 KiB ✅
Codex long turn finished, work page rows 150 150 0 (0.0%) 150 ✅
Codex idle thread resume WebSocket decoded 65 B 65 B 0 B (0.0%) 200 B ✅
Codex idle thread resume WebSocket messages 1 1 0 (0.0%) 1 ✅
Codex fleet shell snapshot HTTP decoded 45.5 KiB 45.5 KiB 0 B (0.0%) 53.7 KiB ✅
Codex fleet shell snapshot rows 27 27 0 (0.0%) 27 ✅
Codex fleet shell subscribe WebSocket decoded 65 B 65 B 0 B (0.0%) 200 B ✅
Codex fleet shell subscribe WebSocket messages 1 1 0 (0.0%) 1 ✅
Codex fleet shell resume check HTTP decoded 18 B 18 B 0 B (0.0%) 100 B ✅
Codex fleet shell resume check rows 0 0 0 0 ✅
Codex stale shell reload HTTP decoded 45.5 KiB 45.5 KiB 0 B (0.0%) 53.7 KiB ✅
Codex stale shell reload rows 27 27 0 (0.0%) 27 ✅
Codex stale shell resume WebSocket decoded 140 B 140 B 0 B (0.0%) 400 B ✅
Codex stale shell resume WebSocket messages 2 2 0 (0.0%) 2 ✅
Codex stale todo list cold open HTTP decoded 72.4 KiB 72.4 KiB 0 B (0.0%) 83.0 KiB ✅
Codex stale todo list cold open rows 75 75 0 (0.0%) 75 ✅
Codex stale todo list cold open nodes 2 2 0 (0.0%) 4 ✅
Codex stale todo list cold open subagents 2 2 0 (0.0%) 4 ✅
Codex stale todo list cold open plans 1 1 0 (0.0%) 2 ✅
Codex stale todo list cold open providerTurns 1 1 0 (0.0%) 2 ✅
Codex stale todo list cold open checkpoints 1 1 0 (0.0%) 2 ✅
Claude short turn resume WebSocket decoded 12.2 KiB 12.2 KiB 0 B (0.0%) 12.7 KiB ✅
Claude short turn resume WebSocket messages 2 2 0 (0.0%) 2 ✅
Claude long turn running, cold open HTTP decoded 283.3 KiB 283.3 KiB 0 B (0.0%) 302.7 KiB ✅
Claude long turn running, cold open rows 75 75 0 (0.0%) 75 ✅
Claude long turn running, history page HTTP decoded 93.4 KiB 93.4 KiB 0 B (0.0%) 107.4 KiB ✅
Claude long turn running, history page rows 75 75 0 (0.0%) 75 ✅
Claude long turn finished, cold open HTTP decoded 384.1 KiB 384.1 KiB 0 B (0.0%) 410.2 KiB ✅
Claude long turn finished, cold open rows 151 151 0 (0.0%) 151 ✅
Claude long turn finished, socket cold open WebSocket decoded 384.3 KiB 384.3 KiB 0 B (0.0%) 410.2 KiB ✅
Claude long turn finished, socket cold open WebSocket messages 2 2 0 (0.0%) 2 ✅
Claude long turn finished, socket cold open rows 151 151 0 (0.0%) 151 ✅
Claude long turn finished, work page HTTP decoded 188.1 KiB 188.1 KiB 0 B (0.0%) 200.2 KiB ✅
Claude long turn finished, work page rows 150 150 0 (0.0%) 150 ✅
Claude idle thread resume WebSocket decoded 66 B 66 B 0 B (0.0%) 200 B ✅
Claude idle thread resume WebSocket messages 1 1 0 (0.0%) 1 ✅
Claude fleet shell snapshot HTTP decoded 46.1 KiB 46.1 KiB 0 B (0.0%) 53.7 KiB ✅
Claude fleet shell snapshot rows 27 27 0 (0.0%) 27 ✅
Claude fleet shell subscribe WebSocket decoded 66 B 66 B 0 B (0.0%) 200 B ✅
Claude fleet shell subscribe WebSocket messages 1 1 0 (0.0%) 1 ✅
Claude fleet shell resume check HTTP decoded 18 B 18 B 0 B (0.0%) 100 B ✅
Claude fleet shell resume check rows 0 0 0 0 ✅
Claude stale shell reload HTTP decoded 46.1 KiB 46.1 KiB 0 B (0.0%) 53.7 KiB ✅
Claude stale shell reload rows 27 27 0 (0.0%) 27 ✅
Claude stale shell resume WebSocket decoded 140 B 140 B 0 B (0.0%) 400 B ✅
Claude stale shell resume WebSocket messages 2 2 0 (0.0%) 2 ✅
Claude stale todo list cold open HTTP decoded 73.4 KiB 73.4 KiB 0 B (0.0%) 83.0 KiB ✅
Claude stale todo list cold open rows 75 75 0 (0.0%) 75 ✅
Claude stale todo list cold open nodes 2 2 0 (0.0%) 4 ✅
Claude stale todo list cold open subagents 2 2 0 (0.0%) 4 ✅
Claude stale todo list cold open plans 1 1 0 (0.0%) 2 ✅
Claude stale todo list cold open providerTurns 1 1 0 (0.0%) 2 ✅
Claude stale todo list cold open checkpoints 1 1 0 (0.0%) 2 ✅

Baseline: a862581 · PR result: 4e89ab9 · Source CI: success

Scenario and decoded snapshot size

10 historical turns, 5 command tools per turn, 878.9 KiB retained MCP result per historical turn, and a 1.05 MiB retained result in the measured turn.
Bounded compact HTTP snapshot with live cursor.

  • Codex decoded thread snapshot: 88.2 KiB
  • Claude decoded thread snapshot: 88.4 KiB

Updated in place by a trusted workflow. PR artifacts are strictly validated and never executed.

Merged via the queue into main with commit d46a10f Nov 20, 2026
91 checks passed
@yaacovcorcos
yaacovcorcos deleted the claude/scient-agent-omp19-20261120 branch November 20, 2026 01:32
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

size:M vouch:trusted Author is trusted by repo permissions or the VOUCHED list.

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant