feat: qualify Pi runtime cutover foundation

This commit is contained in:
2026-08-22 18:35:13 +08:00
parent fba68e86d9
commit 2bc423ebc5
8 changed files with 3182 additions and 2 deletions

View File

@@ -0,0 +1,79 @@
# Task: Implement PI-000 Pi runtime qualification gate
## Identity
- Task ID: 20260822-pi-runtime-qualification-c52e8a14
- Mode: Feature
- Branch: codex/20260822-pi-runtime-qualification-c52e8a14-pi-runtime-qualification
- Worktree: D:\Datas\OthersProjects\makelore-pi-runtime-qualification-c52e8a14
- Base commit: fba68e86d93c66d2c50f53d215de76a7c69c97a7
- Owner: codex
- Status: Completed — Done with explicit waivers
## Scope
- Implement the `PI-000 Runtime qualification gate` defined by the accepted Pi runtime cutover specification.
- Pin the selected Pi runtime package exactly and keep the frozen lockfile reproducible.
- Add a minimal, standalone RPC qualification probe and focused tests for runtime discovery, JSONL framing, session reopen, abort/settled handling, dependency closure, process cleanup, and measurement reporting.
- Produce task-scoped evidence for the supported platform/artifact/provider samples that can actually be run from this worktree.
- Do not change Chat, Host API routes, Renderer behavior, OpenCode migration paths, or final packaged runtime wiring in this task.
## Intent And Constraints
- Candidate runtime: `@earendil-works/pi-coding-agent@0.84.2`, exact version only. Its npm package, CLI entry, Node engine, upstream tag, and commit must agree before it is accepted.
- The intended packaged seam is Electron's executable with `ELECTRON_RUN_AS_NODE=1` and Pi's RPC CLI; the gate must disprove this seam if packaging, concurrency, provider compatibility, resource discovery, or latency budgets fail.
- Required protocol coverage is the current four-protocol Makelore contract: `openai-completions`, `openai-responses`, `anthropic-messages`, and `openrouter`, including base URL, headers, credential injection, model selection, and image-input capability where applicable.
- Windows/Linux and credential-free samples establish the Phase-0 platform evidence. They do not validate real-provider compatibility, concurrency, abort isolation, credential/model isolation, or image-input behavior; the user explicitly waived those remaining `QG-004`/`QG-005` checks on 2026-08-22 and accepted the resulting risk.
- The user explicitly waived macOS x64/arm64 from PI-000 on 2026-08-22 and accepted the risk of discovering macOS packaging/native/resource/performance defects later. This is a deferral to PI-150, not a macOS Pass or removal of final cross-platform release validation.
- The real Provider waiver is not a Pass. Real authentication, endpoint/proxy/rate-limit/provider variation, base URL/header/environment credential/model/image compatibility, true provider concurrency or serialization, and cross-worker abort/event/session/model/credential isolation remain unverified.
- No OpenCode fallback, compatibility layer, RPC/SDK dual path, speculative migration framework, or unrelated refactor is allowed.
- If this gate fails, implementation stops at PI-000 and the planner must amend the specification/ADR to the SDK-in-Electron-utility-process alternative.
## Outcome
- Implemented the exact Pi dependency pin, frozen lockfile update, standalone qualification probe, temporary production-closure staging, controlled Electron Builder `app.asar` probe, real-provider fixture path, four-protocol loopback contract probe, and focused unit coverage.
- Corrected the controlled packaged-closure inspector before macOS execution: the packaged resource directory is now computed once and passed into the artifact process, so a macOS executable reads `Contents/Resources/app.asar` instead of the invalid `Contents/MacOS/resources/app.asar` path.
- Windows x64 and WSL2 Ubuntu 24.04 Linux x64 workspace, staged, and controlled packaged samples are a partial pass. Runtime identity, JSONL RPC, session persistence/reopen, deterministic failure settlement, local worker isolation, shell abort isolation, complete artifact dependency/resource resolution, startup budgets, RSS reporting, and clean exit were confirmed.
- Windows and Linux packaged loopback samples passed request/SSE/image/header/credential/model contracts for `openai-completions`, `openai-responses`, `anthropic-messages`, and `openrouter`, including two-worker overlap and one-worker abort isolation. This is not real Provider Account evidence.
- The planner audited every PI-000 Required work, Acceptance item, and stop condition after the two user decisions. No unwaived Phase-0 item remains incomplete and no confirmed failure exists, so PI-000 is **Done with explicit waivers**, not an all-gates Pass. The new Ready Frontier is `PI-010`, `PI-020`, and `PI-030`.
## Verification
- Project-docs preflight and concurrent/planning gates passed in this isolated feature worktree.
- `@earendil-works/pi-coding-agent@0.84.2` npm metadata and upstream `v0.84.2` tag both resolve to commit `914cf1472e715297caa30db4b9535d534a9eb718`.
- Electron `43.4.0` runs Node `24.18.1` through `ELECTRON_RUN_AS_NODE=1`, satisfying Pi's declared Node `>=22.19.0` engine floor.
- `pnpm exec vitest run tests/unit/pi-runtime-probe.test.ts`: 9 tests passed, including macOS packaged-resource path coverage, aggregate-evidence wording, and explicit real-Provider waiver reporting.
- `pnpm exec eslint scripts/probe-pi-runtime.mjs scripts/probe-pi-packaged-runtime.mjs scripts/probe-pi-provider-contracts.mjs tests/unit/pi-runtime-probe.test.ts`: passed.
- `pnpm run typecheck`: passed.
- `pnpm run build:vite`: passed; existing chunk-size and mixed static/dynamic import warnings remain unchanged and are outside PI-000.
- `node scripts/probe-pi-runtime.mjs --stage --samples 5 --timeout-ms 10000`: partial pass on Windows x64. Cold ready p95 681 ms, warm ready p95 674 ms, RSS p95 125780 KiB, exit p95 16 ms; 130 applicable shrinkwrap packages present and 9 platform-specific entries skipped.
- `node scripts/probe-pi-packaged-runtime.mjs --samples 5 --timeout-ms 10000`: partial pass on a controlled Windows x64 Electron Builder `app.asar` artifact. Cold ready p95 682 ms, warm ready p95 663 ms, RSS p95 141320 KiB, exit p95 18 ms; all 130 applicable packages and six runtime assets resolve from the artifact, with one same-version dependency flattened by Electron Builder.
- `node scripts/probe-pi-provider-contracts.mjs ...packaged-runtime-paths...`: passed all four loopback HTTP/SSE contracts through the packaged executable. Provider-shaped overlap was 304/309/290/297 ms; every protocol passed image/header/credential/model/path checks and one-worker abort isolation. Scope was `127.0.0.1` only and no real Provider Account was contacted.
- Frozen install in an isolated WSL2 Ubuntu 24.04 x64 snapshot used Node `22.22.2` and the exact repository-pinned `pnpm@10.33.4`; native dependencies installed successfully.
- Linux workspace 5+5: cold ready p95 853 ms, warm ready p95 437 ms, RSS p95 170388 KiB, exit p95 8 ms.
- Linux staged 5+5: cold ready p95 428 ms, warm ready p95 440 ms, RSS p95 175956 KiB, exit p95 8 ms; all 131 Linux-applicable packages present and 8 platform-specific entries skipped.
- Linux controlled packaged 5+5: cold ready p95 545 ms, warm ready p95 569 ms, RSS p95 181928 KiB, exit p95 7 ms; all 131 applicable packages, six runtime assets, and unpacked native assets resolve from the artifact.
- Linux packaged loopback: all four contracts passed with 292/290/277/288 ms overlap, correct path/header/credential/model/image serialization, and isolated abort behavior.
- Post-fix `node scripts/probe-pi-packaged-runtime.mjs --samples 1 --timeout-ms 10000`: Windows controlled packaged smoke passed through the explicitly supplied resource directory; cold/warm ready were 650/641 ms and no dependency or asset was missing.
- Post-Provider-waiver `node scripts/probe-pi-runtime.mjs --samples 1 --timeout-ms 10000`: passed with cold/warm ready 650/588 ms. All four `providerMatrix.realTurnVerified` values remained `false`; real Provider evidence moved out of `missingEvidence` into explicit `waivers`, and macOS remained deferred to PI-150. The per-run `decision` remains `incomplete` because one local sample cannot issue the aggregate PI-000 decision.
- Post-probe process queries on Windows and Linux: zero controlled artifact processes and zero Pi RPC-looking processes remained.
- Execution-environment inventory found no repository CI definition, macOS runner/SSH host, GitLab runner CLI, or Provider credential environment variable. The internal GitLab remote alone cannot produce the deferred macOS evidence or optional real-Provider diagnostics; neither absence remains a PI-000 blocker after the explicit user decisions.
- The evidence gate table was reconciled with the accepted Spec's normative numbering: QG-002 RPC lifecycle, QG-003 dependency closure, QG-004 Provider Accounts, QG-005 real-provider concurrency, QG-006 metrics, and QG-007 stop-and-amend condition.
- The planner synchronized the accepted Spec and ticket graph: 134 unique requirements across 19 traceability groups, 18 tickets, 31 edges, and an acyclic topology remain intact; recomputation produced exactly `PI-010`, `PI-020`, and `PI-030` as the new Ready Frontier.
- Detailed evidence: `.project-docs/50-evidence/topics/20260822-pi-runtime-qualification-c52e8a14__pi-runtime-qualification.md`.
## Follow-ups
- Real Provider fixtures may still be run as optional diagnostics, but they are no longer PI-000 or PI-150 acceptance requirements. Never backfill QG-004/QG-005 as Pass without actual evidence, and do not place secrets in fixtures or task records.
- At PI-150, run macOS x64 and macOS arm64 workspace/staged/final packaged samples, packaged loopback, closure/native/resource checks, and metrics before claiming cross-platform release readiness. Use separate frozen installs and artifacts for each architecture.
- If Linux qualification must cover a native desktop distribution or release package rather than the accepted controlled `dir` boundary, repeat the Linux probe outside WSL2; the current sample is explicitly Ubuntu 24.04 userspace on a WSL2 kernel.
- Any later confirmed runtime, packaging, concurrency, isolation, or performance failure is new evidence and must return to architecture review; this waiver records accepted uncertainty, not permission to misreport a known failure.
## Promotion Candidates
- Target: the planner's Pi runtime cutover ticket graph / ready frontier.
Proposal: mark PI-000 Done with explicit waivers and advance the Ready Frontier to `PI-010`, `PI-020`, and `PI-030`; keep macOS explicitly deferred to PI-150 and QG-004/QG-005 explicitly waived rather than passed.
Evidence: task-scoped PI-000 qualification evidence record.
Future impact: downstream implementation proceeds while carrying explicit real-Provider and deferred macOS uncertainty; loopback and packaged evidence remain the proven boundary.
Semantic conflicts: none after the planner's user-authorized Spec/ticket amendment.
Human confirmation required: already supplied by the user's 2026-08-22 waiver decisions; a later confirmed failure requires a new architecture decision.

View File

@@ -0,0 +1,232 @@
# PI-000 Pi Runtime Qualification Evidence
## Scope And Decision State
- Task: `20260822-pi-runtime-qualification-c52e8a14`
- Candidate: `@earendil-works/pi-coding-agent@0.84.2`
- Current decision: **PI-000 Done with explicit waivers**. This is not an all-gates Pass.
- Ready frontier consequence: `PI-010`, `PI-020`, and `PI-030` are unlocked by the planner's audited 18-node/31-edge DAG.
- macOS Phase-0 decision: **Explicitly Waived / Deferred by user on 2026-08-22**. This is not a Pass. macOS x64/arm64 qualification moves to `PI-150` final cross-platform release validation.
- Real Provider decision: `QG-004` and `QG-005` are **Explicitly Waived / Accepted Risk by user on 2026-08-22**, not Pass. They no longer block PI-000 or PI-150.
- Accepted unverified risk includes real authentication, endpoint/proxy/rate-limit/provider-specific response variation; base URL, header, environment credential, model and image compatibility; real provider concurrency/serialization; and cross-worker abort/event/session/model/credential isolation.
- This record summarizes reproducible evidence. Raw generated JSON reports, temporary staged runtimes, and controlled directory artifacts are intentionally not committed.
## Confirmed Upstream Identity
- npm package version: `0.84.2`
- npm CLI entry: `dist/cli.js`
- declared Node engine: `>=22.19.0`
- upstream tag: `v0.84.2`
- upstream commit and npm `gitHead`: `914cf1472e715297caa30db4b9535d534a9eb718`
- package pin in Makelore: exact `0.84.2`, without caret or tilde
- project package manager used for the lockfile: the repository-pinned `pnpm@10.33.4`
The selected Electron `43.4.0` executable exposes Node `24.18.1` when launched with `ELECTRON_RUN_AS_NODE=1`, which satisfies Pi's declared engine floor.
## Windows x64 Qualification Sample
Environment:
- OS: Windows x64, release `10.0.26200`
- Runtime seam: Electron executable plus `ELECTRON_RUN_AS_NODE=1`
- Artifact: a temporary production-only staging directory installed from Pi's published `npm-shrinkwrap.json`
- Startup samples: 5 cold and 5 warm
- Cold definition: fresh Pi config, session, and project directories for each process
- Warm definition: shared primed Pi config directory with fresh session and project directories for each process
Measured result:
| Measurement | Samples | p50 | p95 | max | Budget |
|---|---:|---:|---:|---:|---:|
| Cold `get_state` ready | 5 | 654 ms | 681 ms | 681 ms | 3000 ms |
| Warm `get_state` ready | 5 | 661 ms | 674 ms | 674 ms | 1500 ms |
| Worker RSS | 10 | 122780 KiB | 125780 KiB | 125780 KiB | report-only |
| RPC process exit | 10 | 14 ms | 16 ms | 16 ms | deterministic |
All measured processes exited by closing stdin. No SIGTERM, force kill, or retained child was needed in the successful sample.
The deterministic provider-failure sample used a closed loopback endpoint, disabled automatic retry, and reached `prompt` accepted followed by `agent_settled` with an assistant `error` stop reason in 42 ms. This is below the 10-second deterministic-failure limit and does not claim provider success.
## Controlled Electron Builder Artifact Sample
PI-000 also built a Windows x64 Electron Builder `dir` artifact from the same temporary Pi production closure. This is a controlled packaged-boundary probe, not Makelore's final product artifact and not the permanent `PI-030`/`PI-150` bundler wiring. It deliberately does not inherit the product's OpenCode, Python, UV, signing, installer, or other `electron-builder.yml` hooks.
The packaged boundary was:
- executable: `MakelorePiProbe.exe`
- archive: `resources/app.asar`
- native extraction: `resources/app.asar.unpacked`
- Pi CLI: `resources/app.asar/dist/cli.js`
- process seam: packaged executable plus `ELECTRON_RUN_AS_NODE=1`
Artifact resolution confirmed Pi `0.84.2`, its CLI, all 130 Windows-applicable production packages, all six runtime WASM/native assets, and native `.node` extraction. The nine non-Windows lock entries were skipped by their declared platform constraints. Electron Builder flattened `@types/retry@0.12.0` from `p-retry/node_modules/@types/retry` to the artifact root `node_modules/@types/retry`; the verifier checks the same version at the resolved artifact location instead of requiring npm's original physical layout.
Measured result from 5 cold and 5 warm packaged samples:
| Measurement | Samples | p50 | p95 | max | Budget |
|---|---:|---:|---:|---:|---:|
| Cold `get_state` ready | 5 | 668 ms | 682 ms | 682 ms | 3000 ms |
| Warm `get_state` ready | 5 | 660 ms | 663 ms | 663 ms | 1500 ms |
| Worker RSS | 10 | 138404 KiB | 141320 KiB | 141320 KiB | report-only |
| RPC process exit | 10 | 16 ms | 18 ms | 18 ms | deterministic |
The packaged session, deterministic failure settlement, reopen, local shell overlap, one-worker shell abort isolation, and stdin-close exit checks also passed. The deterministic failure settled in 36 ms; the overlapping local shell sample completed in 2098 ms. A post-run process query found zero controlled artifact processes and zero Pi RPC-looking processes.
## Linux x64 Qualification Sample
The Linux run used an isolated repository snapshot under WSL2 Ubuntu 24.04 x64. The snapshot installed the frozen lockfile with Node `22.22.2` and the exact repository-pinned `pnpm@10.33.4`; Pi then ran through Electron `43.4.0` and its embedded Node `24.18.1`. This exercises the Linux executable, Linux native dependency selection, asar layout, and process behavior. It is not evidence for a native compositor, AppImage/deb installer, or a non-WSL Linux kernel.
Workspace 5+5 result:
| Measurement | Samples | p50 | p95 | max | Budget |
|---|---:|---:|---:|---:|---:|
| Cold `get_state` ready | 5 | 426 ms | 853 ms | 853 ms | 3000 ms |
| Warm `get_state` ready | 5 | 432 ms | 437 ms | 437 ms | 1500 ms |
| Worker RSS | 10 | 168228 KiB | 170388 KiB | 170388 KiB | report-only |
| RPC process exit | 10 | 7 ms | 8 ms | 8 ms | deterministic |
Temporary production-closure 5+5 result:
| Measurement | Samples | p50 | p95 | max | Budget |
|---|---:|---:|---:|---:|---:|
| Cold `get_state` ready | 5 | 426 ms | 428 ms | 428 ms | 3000 ms |
| Warm `get_state` ready | 5 | 423 ms | 440 ms | 440 ms | 1500 ms |
| Worker RSS | 10 | 170380 KiB | 175956 KiB | 175956 KiB | report-only |
| RPC process exit | 10 | 7 ms | 8 ms | 8 ms | deterministic |
Controlled Electron Builder `dir`/`app.asar` 5+5 result:
| Measurement | Samples | p50 | p95 | max | Budget |
|---|---:|---:|---:|---:|---:|
| Cold `get_state` ready | 5 | 526 ms | 545 ms | 545 ms | 3000 ms |
| Warm `get_state` ready | 5 | 545 ms | 569 ms | 569 ms | 1500 ms |
| Worker RSS | 10 | 181412 KiB | 181928 KiB | 181928 KiB | report-only |
| RPC process exit | 10 | 7 ms | 7 ms | 7 ms | deterministic |
The Linux shrinkwrap has 131 applicable packages and 8 OS/CPU-skipped entries. The controlled artifact resolved all applicable packages, the Pi CLI, all six tracked runtime assets, and every native asset that must live under `app.asar.unpacked`; no package or asset was missing. Electron Builder applied the same-version `@types/retry@0.12.0` flattening seen on Windows. Session persistence/reopen, deterministic failure settlement, local overlap, shell abort isolation, and stdin-close cleanup passed in all three boundaries. A final process listing contained only WSL init and the listing command itself—no packaged Pi or RPC worker remained.
## Runtime And Session Behaviors Confirmed On Windows
- Strict LF-only JSONL framing, including payloads containing U+2028 and U+2029.
- `get_state` response and request id correlation.
- Persistent session file creation after a settled assistant turn.
- Reopen by session file with the same stable session id.
- Session history remained readable after reopen; the sample contained six entries.
- Direct RPC shell execution returned the expected output.
- Idle `abort` was accepted.
- Two workers had distinct session ids.
- Two direct shell operations overlapped; total elapsed time was 2112 ms for a 4-second left command and 2-second right command.
- `abort_bash` cancelled the left worker without cancelling or corrupting the right worker.
- Resource discovery was explicitly disabled for the qualification sample with `--no-extensions`, `--no-skills`, `--no-prompt-templates`, `--no-themes`, `--no-context-files`, and `--no-approve`; startup ran with `--offline` and `PI_TELEMETRY=0`.
The two-worker result proves local process/session/shell isolation only. It is not evidence of two overlapping real-provider turns.
## Production Closure Evidence
Pi's published production shrinkwrap contains 139 dependency package entries:
- 130 entries are applicable and present on Windows x64.
- 9 entries are correctly skipped by OS/CPU constraints.
- Required image-processing asset is present: `@silvia-odwyer/photon-node/photon_rs_bg.wasm`.
- Applicable clipboard native binary is present: `@mariozechner/clipboard-win32-x64-msvc`.
- Pi TUI Windows x64/arm64 prebuilds and its packaged Darwin prebuilds are present in the staged closure.
Published-package caveat: the shrinkwrap is production-only, but the published package still lists ten development dependencies. npm 11 validates those development entries before applying `--omit=dev`, so an unmodified `npm ci --omit=dev` rejects the package as out of sync. The qualification staging copy removes `devDependencies` from only the temporary copied package root and then installs the unchanged published shrinkwrap. This must be considered when `PI-030` defines the permanent staging mechanism.
## Provider Contract Matrix
Static Pi contract coverage and a packaged-runtime loopback HTTP/SSE contract sample are confirmed for all current Makelore protocol identifiers:
| Makelore protocol | Pi configuration | Loopback endpoint | Windows / Linux overlap | Real turn |
|---|---|---|---:|---|
| `openai-completions` | `api: openai-completions` | `/v1/chat/completions` | 304 / 292 ms | not run — waived/unverified |
| `openai-responses` | `api: openai-responses` | `/v1/responses` | 309 / 290 ms | not run — waived/unverified |
| `anthropic-messages` | `api: anthropic-messages` | `/v1/messages` | 290 / 277 ms | not run — waived/unverified |
| `openrouter` | `api: openai-completions` plus `thinkingFormat/sessionAffinityFormat: openrouter` | `/v1/chat/completions` | 297 / 288 ms | not run — waived/unverified |
For each protocol, the packaged runtime sent four POST requests through `127.0.0.1`, used the selected model, included the custom header, injected the environment-only credential into the protocol-appropriate header (`Authorization` or `x-api-key`), and carried the PNG image in request history. Two independent workers settled successfully with `stop`, their provider-shaped turns overlapped, aborting the next turn in one worker produced `aborted`, and the unaffected worker still settled with `stop`. All workers exited by closing stdin.
The Anthropic contract has a meaningful base URL difference: its SDK appends `/v1/messages`, so the configured base URL is the provider root. The OpenAI-style clients receive a base URL ending in `/v1` and append `chat/completions` or `responses`.
This loopback result proves Pi's request serialization, SSE parsing, image/header/model/credential plumbing, worker overlap, and abort isolation without external cost. It does **not** qualify any real Provider Account, provider authentication, proxy behavior, rate limits, or provider-specific response variation.
Pi's custom provider contract supports base URL, environment-resolved API key, custom headers, model id, and text/image input declaration. The probe accepts a credential-free fixture file shaped as follows; the secret remains only in the named environment variable:
```json
{
"id": "qualification-provider",
"apiProtocol": "openai-responses",
"baseUrl": "https://provider.example/v1",
"apiKeyEnv": "QUALIFICATION_PROVIDER_KEY",
"headers": { "X-Provider-Header": "literal-or-$ENV_REFERENCE" },
"model": {
"id": "provider-model-id",
"input": ["text", "image"],
"contextWindow": 128000,
"maxTokens": 8192
}
}
```
Run a real sample with:
```text
pnpm run probe:pi-runtime -- --provider-fixture <fixture.json> [--image <image-path>]
```
The real-provider path is implemented to require two successful overlapping turns with distinct sessions, then abort a new turn in one worker while the other worker settles successfully. It reports only provider/model ids, environment-variable name, and header names; it does not print credential values.
No recognized provider credential environment variable was available in the implementation environment, so no real-provider request was sent and no provider cost was incurred.
Reproduce the local contract sample against an existing controlled artifact with:
```text
pnpm run probe:pi-provider-contracts -- --electron-executable <artifact-executable> --cli-path <artifact-app.asar-cli>
```
## Deferred macOS Execution Handoff For PI-150
A pre-execution portability review found and fixed one supported-path defect: the artifact-side closure inspector derived `resources` from `dirname(process.execPath)`, which is correct on Windows/Linux but points to `Contents/MacOS/resources` on macOS. The probe now computes the platform resource directory once and passes it into the packaged process, preserving the actual macOS location `Contents/Resources/app.asar`. A focused path test and a real Windows controlled packaged regression pass verify the shared logic. The user explicitly deferred actual macOS execution from PI-000 to PI-150 on 2026-08-22, so no macOS Pass is claimed and the late-discovery risk remains.
Use a separate checkout and frozen install for each target architecture. Before installing, `node -p "process.platform + ' ' + process.arch + ' ' + process.version"` must report `darwin x64` or `darwin arm64` for the sample being collected. Do not reuse one architecture's `node_modules` or Electron download for the other architecture.
Run this sequence in each architecture-specific checkout:
```text
pnpm install --frozen-lockfile
node scripts/probe-pi-runtime.mjs --samples 5 --timeout-ms 10000 --report release/pi-runtime-macos-workspace.json
pnpm run probe:pi-runtime -- --samples 5 --timeout-ms 10000 --report release/pi-runtime-macos-staged.json
pnpm run probe:pi-runtime:packaged -- --samples 5 --timeout-ms 10000 --report release/pi-runtime-macos-packaged.json
```
Read `artifact.executable` and `artifact.appAsar` from the packaged JSON report. The Pi CLI path is `<artifact.appAsar>/dist/cli.js`. Run the credential-free packaged four-protocol contract sample before replacing the controlled artifact:
```text
pnpm run probe:pi-provider-contracts -- --electron-executable <artifact.executable> --cli-path <artifact.appAsar>/dist/cli.js --artifact-label controlled-electron-builder-dir-app-asar --report release/pi-provider-contract-macos.json
```
For an optional, non-gating real Provider diagnostic, create one credential-free fixture per protocol being sampled and keep each secret only in the environment variable named by that fixture's `apiKeyEnv`. Reuse the existing packaged artifact through the runtime override so the run exercises the packaged Electron seam without rebuilding it for every protocol:
```text
node scripts/probe-pi-runtime.mjs --samples 5 --timeout-ms 10000 --electron-executable <artifact.executable> --cli-path <artifact.appAsar>/dist/cli.js --artifact-label controlled-electron-builder-dir-app-asar --provider-fixture <protocol-fixture.json> --image <qualification-image> --report release/pi-provider-<protocol>-macos.json
```
The user waived requiring these commands for `openai-completions`, `openai-responses`, `anthropic-messages`, and `openrouter`. If an optional diagnostic is run, preserve only redacted metrics and protocol/provider/model identifiers; do not commit fixtures containing secrets, generated reports, controlled artifacts, caches, or user data. A final process query must show no packaged probe or Pi RPC worker remains.
## Gate Status
| Gate | Status | Evidence or blocker |
|---|---|---|
| QG-001 exact identity and pin | Pass | exact npm/tag/commit/engine/CLI and lockfile pin |
| QG-002 packaged Electron Node RPC lifecycle | Pass | Phase-0 scope is Windows x64 + Linux x64 by explicit user decision; both controlled artifacts spawn, become ready, open session, accept prompt, abort, settle, reopen, and exit cleanly. macOS is deferred to PI-150, not passed. |
| QG-003 production dependency closure | Pass | Phase-0 controlled Windows/Linux artifacts resolve all 130/131 applicable packages and six runtime assets. Final product artifacts and deferred macOS closure belong to PI-150. |
| QG-004 Provider Account matrix | Explicitly Waived / Accepted Risk | four-protocol loopback request/SSE/image contracts pass; real accounts, credentials, endpoints, proxies, authentication, rate-limit and provider-specific behavior remain unverified by user decision on 2026-08-22 |
| QG-005 real-provider worker concurrency | Explicitly Waived / Accepted Risk | local shell and loopback protocol turns overlap with isolated abort/model/credential state; real provider overlap/serialization and abort/event/session/model/credential isolation remain unverified by user decision on 2026-08-22 |
| QG-006 performance and cleanup metrics | Pass | Phase-0 Windows/Linux cold/warm/RSS/exit sample counts and p50/p95/max are recorded and within budgets. macOS metrics are explicitly deferred to PI-150. |
| QG-007 stop-and-amend condition | Pending / not triggered | no confirmed packaging, Provider, or budget failure yet; any mandatory-gate failure stops migration and requires the specified utility-process SDK amendment |
## Remaining And Optional Evidence
1. Real Provider fixtures are optional diagnostics after the explicit waiver. Their absence does not block PI-000 or PI-150, but the current record must continue to report every real turn as unverified; a later confirmed failure is new architecture evidence, not something the waiver converts into a Pass.
2. At PI-150, run the deferred macOS x64/arm64 workspace, staged, final packaged, loopback, closure/native/resource, and metrics matrix. Missing either architecture blocks a cross-platform release-ready claim even though it no longer blocks PI-000.
3. If the accepted Linux definition requires a native desktop distribution or release package, repeat the Linux sample outside WSL2 at PI-150; the current Phase-0 evidence is deliberately labeled as Ubuntu 24.04 userspace on WSL2 with a controlled `dir` artifact.

View File

@@ -37,6 +37,9 @@
"build": "pnpm run build:vite && node scripts/bundle-opencode.mjs && node scripts/run-electron-builder.mjs",
"build:vite": "node --max-old-space-size=6144 ./node_modules/vite/bin/vite.js build",
"perf:budget": "node scripts/check-performance-budget.mjs",
"probe:pi-runtime": "node scripts/probe-pi-runtime.mjs --stage",
"probe:pi-runtime:packaged": "node scripts/probe-pi-packaged-runtime.mjs",
"probe:pi-provider-contracts": "node scripts/probe-pi-provider-contracts.mjs",
"bundle:opencode": "node scripts/bundle-opencode.mjs",
"lint": "eslint . --fix",
"lint:check": "eslint .",
@@ -84,6 +87,7 @@
"dependencies": {
"@napi-rs/canvas": "0.1.100",
"@monaco-editor/react": "^4.7.0",
"@earendil-works/pi-coding-agent": "0.84.2",
"@playwright/mcp": "0.0.78",
"adm-zip": "0.5.16",
"diff": "^9.0.0",

972
pnpm-lock.yaml generated

File diff suppressed because it is too large Load Diff

View File

@@ -0,0 +1,322 @@
import { spawn } from 'node:child_process';
import { mkdtemp, mkdir, readFile, readdir, rm, stat, writeFile } from 'node:fs/promises';
import { arch, platform, tmpdir } from 'node:os';
import { dirname, join, relative, resolve, sep } from 'node:path';
import { pathToFileURL } from 'node:url';
import { Arch, Platform, build } from 'electron-builder';
import {
parseProbeArgs,
resolvePiRuntime,
runProbe,
stagePiRuntime,
validatePiIdentity,
validateStagedClosure,
} from './probe-pi-runtime.mjs';
const PRODUCT_NAME = 'MakelorePiProbe';
const ARTIFACT_LABEL = 'controlled-electron-builder-dir-app-asar';
async function pathExists(path) {
try {
await stat(path);
return true;
} catch {
return false;
}
}
function currentTarget() {
let platformTarget;
if (platform() === 'win32') platformTarget = Platform.WINDOWS;
else if (platform() === 'darwin') platformTarget = Platform.MAC;
else if (platform() === 'linux') platformTarget = Platform.LINUX;
else throw new Error(`Unsupported packaging platform: ${platform()}`);
const architecture = Arch[arch()];
if (architecture == null) throw new Error(`Unsupported packaging architecture: ${arch()}`);
return { platformTarget, architecture };
}
async function findPackagedExecutable(root) {
const matches = [];
const visit = async (directory) => {
for (const entry of await readdir(directory, { withFileTypes: true })) {
const path = join(directory, entry.name);
if (entry.isDirectory()) await visit(path);
else if (
(platform() === 'win32' && entry.name === `${PRODUCT_NAME}.exe`)
|| (platform() === 'darwin'
&& entry.name === PRODUCT_NAME
&& directory.endsWith(`${sep}Contents${sep}MacOS`))
|| (platform() === 'linux' && entry.name === PRODUCT_NAME)
) matches.push(path);
}
};
await visit(root);
if (matches.length !== 1) {
throw new Error(`Expected one packaged ${PRODUCT_NAME} executable, found ${matches.length}`);
}
return matches[0];
}
export function packagedResourcesDirectory(executable, currentPlatform = platform()) {
return currentPlatform === 'darwin'
? resolve(executable, '..', '..', 'Resources')
: join(dirname(executable), 'resources');
}
function runCommand(executable, args, options = {}) {
return new Promise((resolvePromise, reject) => {
const child = spawn(executable, args, {
cwd: options.cwd,
env: options.env ?? process.env,
stdio: ['ignore', 'pipe', 'pipe'],
windowsHide: true,
});
let stdout = '';
let stderr = '';
child.stdout.on('data', (chunk) => { stdout += chunk.toString(); });
child.stderr.on('data', (chunk) => { stderr += chunk.toString(); });
child.once('error', reject);
child.once('exit', (code, signal) => {
if (code === 0) resolvePromise({ stdout, stderr });
else reject(new Error(
`${executable} exited with code ${code ?? 'null'} signal ${signal ?? 'none'}: ${stderr || stdout}`,
));
});
});
}
async function inspectPackagedClosure(executable, resourcesDirectory, assets) {
const script = String.raw`
const fs = require('node:fs');
const path = require('node:path');
const resources = process.env.PI_PROBE_RESOURCES_DIRECTORY;
if (!resources) throw new Error('PI_PROBE_RESOURCES_DIRECTORY is required');
const root = path.join(resources, 'app.asar');
const unpackedRoot = path.join(resources, 'app.asar.unpacked');
const packageJson = JSON.parse(fs.readFileSync(path.join(root, 'package.json'), 'utf8'));
const shrinkwrap = JSON.parse(fs.readFileSync(path.join(root, 'npm-shrinkwrap.json'), 'utf8'));
const assets = JSON.parse(process.env.PI_PROBE_ASSETS_JSON);
const libc = process.platform === 'linux'
? (process.report?.getReport?.().header?.glibcVersionRuntime ? 'glibc' : 'musl')
: undefined;
const constraintMatches = (constraints, current) => {
if (!constraints || constraints.length === 0 || !current) return true;
const positive = constraints.filter((value) => !value.startsWith('!'));
const negative = constraints.filter((value) => value.startsWith('!')).map((value) => value.slice(1));
return !negative.includes(current) && (positive.length === 0 || positive.includes(current));
};
const supportsCurrentPlatform = (entry) => constraintMatches(entry.os, process.platform)
&& constraintMatches(entry.cpu, process.arch)
&& constraintMatches(entry.libc, libc);
const missingPackages = [];
const relocatedPackages = [];
let expectedPackages = 0;
let platformSkippedPackages = 0;
for (const [packagePath, entry] of Object.entries(shrinkwrap.packages || {})) {
if (!packagePath || entry.dev) continue;
if (!supportsCurrentPlatform(entry)) {
platformSkippedPackages += 1;
continue;
}
expectedPackages += 1;
const exactPath = path.join(root, ...packagePath.split('/'));
if (fs.existsSync(exactPath)) continue;
const nestedMarker = '/node_modules/';
const packageTail = packagePath.includes(nestedMarker)
? packagePath.slice(packagePath.lastIndexOf(nestedMarker) + nestedMarker.length)
: packagePath.slice('node_modules/'.length);
const nameParts = packageTail.split('/');
const packageName = nameParts[0].startsWith('@') ? nameParts.slice(0, 2).join('/') : nameParts[0];
const flattenedPath = path.join(root, 'node_modules', ...packageName.split('/'));
const flattenedPackageJson = path.join(flattenedPath, 'package.json');
if (fs.existsSync(flattenedPackageJson)) {
const flattenedVersion = JSON.parse(fs.readFileSync(flattenedPackageJson, 'utf8')).version;
if (flattenedVersion === entry.version) {
relocatedPackages.push({ from: packagePath, to: 'node_modules/' + packageName, version: entry.version });
continue;
}
}
missingPackages.push(packagePath);
}
const missingAssets = assets.filter((asset) => !fs.existsSync(path.join(root, ...asset.split('/'))));
const missingUnpackedNativeAssets = assets
.filter((asset) => asset.endsWith('.node'))
.filter((asset) => !fs.existsSync(path.join(unpackedRoot, ...asset.split('/'))));
process.stdout.write(JSON.stringify({
packageName: packageJson.name,
packageVersion: packageJson.version,
cliExists: fs.existsSync(path.join(root, 'dist', 'cli.js')),
lockedPackages: Math.max(0, Object.keys(shrinkwrap.packages || {}).length - 1),
expectedPackages,
platformSkippedPackages,
missingPackages,
relocatedPackages,
assetCount: assets.length,
missingAssets,
missingUnpackedNativeAssets,
}));
`;
const { stdout } = await runCommand(executable, ['-e', script], {
env: {
...process.env,
ELECTRON_RUN_AS_NODE: '1',
PI_PROBE_RESOURCES_DIRECTORY: resourcesDirectory,
PI_PROBE_ASSETS_JSON: JSON.stringify(assets),
},
});
const result = JSON.parse(stdout.trim());
if (
!result.cliExists
|| result.missingPackages.length > 0
|| result.missingAssets.length > 0
|| result.missingUnpackedNativeAssets.length > 0
) {
throw new Error(`Packaged Pi closure is incomplete: ${JSON.stringify(result)}`);
}
return result;
}
function assertControlledOutput(projectRoot, outputDirectory) {
const relativePath = relative(projectRoot, outputDirectory);
if (!relativePath || relativePath.startsWith('..') || resolve(projectRoot, relativePath) !== outputDirectory) {
throw new Error(`Refusing uncontrolled packaged-probe output: ${outputDirectory}`);
}
if (relativePath.split(sep).join('/') !== 'release/pi-runtime-probe') {
throw new Error(`Unexpected packaged-probe output directory: ${outputDirectory}`);
}
}
async function buildControlledArtifact(projectRoot, appDirectory, outputDirectory) {
const electronPackage = JSON.parse(await readFile(join(projectRoot, 'node_modules', 'electron', 'package.json'), 'utf8'));
const { platformTarget, architecture } = currentTarget();
const buildResources = join(appDirectory, 'build-resources');
await mkdir(buildResources);
const builderArtifacts = await build({
// Use the staged Pi root as the project root so this controlled probe does
// not inherit Makelore's production hooks or unrelated extraResources.
projectDir: appDirectory,
targets: platformTarget.createTarget('dir', architecture),
publish: 'never',
config: {
appId: 'app.niancode.desktop.pi-probe',
productName: PRODUCT_NAME,
electronVersion: electronPackage.version,
electronDist: join(projectRoot, 'node_modules', 'electron', 'dist'),
directories: {
output: outputDirectory,
buildResources,
},
files: [
'dist/**/*',
'package.json',
'npm-shrinkwrap.json',
'node_modules/**/*',
],
asar: true,
asarUnpack: ['**/*.node'],
npmRebuild: false,
win: {
executableName: PRODUCT_NAME,
signAndEditExecutable: false,
verifyUpdateCodeSignature: false,
},
mac: {
identity: null,
hardenedRuntime: false,
},
linux: {
executableName: PRODUCT_NAME,
},
},
});
return builderArtifacts;
}
export async function runPackagedProbe(options, projectRoot = process.cwd()) {
const resolvedProjectRoot = resolve(projectRoot);
const outputDirectory = resolve(resolvedProjectRoot, 'release', 'pi-runtime-probe');
assertControlledOutput(resolvedProjectRoot, outputDirectory);
const scratchRoot = await mkdtemp(join(tmpdir(), 'makelore-pi-packaged-probe-'));
const appDirectory = join(scratchRoot, 'app');
try {
const sourceRuntime = await resolvePiRuntime(resolvedProjectRoot);
const identity = await validatePiIdentity(sourceRuntime);
await stagePiRuntime(sourceRuntime, appDirectory);
const stagedClosure = await validateStagedClosure(appDirectory);
await rm(outputDirectory, { recursive: true, force: true, maxRetries: 5, retryDelay: 200 });
await mkdir(dirname(outputDirectory), { recursive: true });
const builderArtifacts = await buildControlledArtifact(
resolvedProjectRoot,
appDirectory,
outputDirectory,
);
const executable = await findPackagedExecutable(outputDirectory);
const resourcesDirectory = packagedResourcesDirectory(executable);
const appAsar = join(resourcesDirectory, 'app.asar');
const cliPath = join(appAsar, 'dist', 'cli.js');
if (!await pathExists(appAsar)) throw new Error(`Packaged app.asar is missing: ${appAsar}`);
const packagedClosure = await inspectPackagedClosure(
executable,
resourcesDirectory,
stagedClosure.assets,
);
const runtime = await runProbe({
...options,
reportPath: undefined,
stage: false,
keepStage: false,
electronExecutablePath: executable,
cliPath,
artifactLabel: ARTIFACT_LABEL,
}, resolvedProjectRoot);
const report = {
schemaVersion: 1,
generatedAt: new Date().toISOString(),
identity,
artifact: {
label: ARTIFACT_LABEL,
outputDirectory,
executable,
appAsar,
appAsarUnpacked: join(resourcesDirectory, 'app.asar.unpacked'),
builderArtifacts,
},
stagedClosure,
packagedClosure,
runtime,
result: runtime.result,
decision: 'incomplete',
};
if (options.reportPath) {
await mkdir(dirname(options.reportPath), { recursive: true });
await writeFile(options.reportPath, `${JSON.stringify(report, null, 2)}\n`);
}
return report;
} finally {
await rm(scratchRoot, { recursive: true, force: true, maxRetries: 5, retryDelay: 200 });
}
}
async function main() {
const options = parseProbeArgs(process.argv.slice(2));
if (options.help) {
process.stdout.write('Usage: node scripts/probe-pi-packaged-runtime.mjs [probe options]\n');
return;
}
if (options.stage || options.keepStage || options.electronExecutablePath) {
throw new Error('Packaged probe owns staging and runtime paths; do not pass staging or runtime override flags');
}
const report = await runPackagedProbe(options);
process.stdout.write(`${JSON.stringify(report, null, 2)}\n`);
if (report.result === 'fail') process.exitCode = 1;
}
const isMain = process.argv[1]
&& pathToFileURL(resolve(process.argv[1])).href === import.meta.url;
if (isMain) {
main().catch((error) => {
process.stderr.write(`${error.stack ?? error.message}\n`);
process.exitCode = 1;
});
}

View File

@@ -0,0 +1,324 @@
import { createServer } from 'node:http';
import { mkdtemp, mkdir, rm, writeFile } from 'node:fs/promises';
import { tmpdir } from 'node:os';
import { dirname, join, resolve } from 'node:path';
import { pathToFileURL } from 'node:url';
import {
MAKELore_PROVIDER_PROTOCOLS,
parseProbeArgs,
resolvePiRuntime,
runProviderQualification,
validatePiIdentity,
} from './probe-pi-runtime.mjs';
const LOCAL_API_KEY_ENV = 'PI_PROBE_LOCAL_CONTRACT_KEY';
const LOCAL_API_KEY = 'pi-local-contract-only';
const CUSTOM_HEADER = 'x-makelore-pi-probe';
const IMAGE_BASE64 = 'iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAQAAAC1HAwCAAAAC0lEQVR42mP8/x8AAusB9Y9Z4N8AAAAASUVORK5CYII=';
function readRequestBody(request) {
return new Promise((resolvePromise, reject) => {
const chunks = [];
request.on('data', (chunk) => chunks.push(chunk));
request.once('end', () => {
try {
resolvePromise(JSON.parse(Buffer.concat(chunks).toString('utf8')));
} catch (error) {
reject(error);
}
});
request.once('error', reject);
});
}
function writeSse(response, event) {
if (response.destroyed || response.writableEnded) return;
response.write(`event: ${event.type}\ndata: ${JSON.stringify(event)}\n\n`);
}
function finishResponse(response, events) {
setTimeout(() => {
if (response.destroyed || response.writableEnded) return;
for (const event of events) writeSse(response, event);
response.end();
}, 250);
}
function respondOpenAiCompletions(response, body) {
const id = `chatcmpl-local-${Date.now()}`;
response.writeHead(200, { 'content-type': 'text/event-stream' });
response.write(`data: ${JSON.stringify({
id,
object: 'chat.completion.chunk',
created: Math.floor(Date.now() / 1000),
model: body.model,
choices: [{ index: 0, delta: { role: 'assistant', content: 'PI_PROVIDER_OK' }, finish_reason: null }],
})}\n\n`);
setTimeout(() => {
if (response.destroyed || response.writableEnded) return;
response.write(`data: ${JSON.stringify({
id,
object: 'chat.completion.chunk',
created: Math.floor(Date.now() / 1000),
model: body.model,
choices: [{ index: 0, delta: {}, finish_reason: 'stop' }],
usage: { prompt_tokens: 1, completion_tokens: 1, total_tokens: 2 },
})}\n\n`);
response.end('data: [DONE]\n\n');
}, 250);
}
function responseEnvelope(id, model, status, output = []) {
return {
id,
object: 'response',
created_at: Math.floor(Date.now() / 1000),
status,
model,
output,
error: null,
incomplete_details: null,
usage: status === 'completed'
? {
input_tokens: 1,
input_tokens_details: { cached_tokens: 0 },
output_tokens: 1,
output_tokens_details: { reasoning_tokens: 0 },
total_tokens: 2,
}
: null,
};
}
function respondOpenAiResponses(response, body) {
const id = `resp_local_${Date.now()}`;
const item = {
id: `msg_local_${Date.now()}`,
type: 'message',
status: 'completed',
role: 'assistant',
content: [{ type: 'output_text', text: 'PI_PROVIDER_OK', annotations: [] }],
};
response.writeHead(200, { 'content-type': 'text/event-stream' });
writeSse(response, { type: 'response.created', response: responseEnvelope(id, body.model, 'in_progress') });
writeSse(response, { type: 'response.output_item.added', output_index: 0, item: { ...item, status: 'in_progress', content: [] } });
writeSse(response, {
type: 'response.output_text.delta',
item_id: item.id,
output_index: 0,
content_index: 0,
delta: 'PI_PROVIDER_OK',
});
finishResponse(response, [
{ type: 'response.output_item.done', output_index: 0, item },
{ type: 'response.completed', response: responseEnvelope(id, body.model, 'completed', [item]) },
]);
}
function respondAnthropic(response, body) {
const id = `msg_local_${Date.now()}`;
response.writeHead(200, { 'content-type': 'text/event-stream' });
writeSse(response, {
type: 'message_start',
message: {
id,
type: 'message',
role: 'assistant',
model: body.model,
content: [],
stop_reason: null,
stop_sequence: null,
usage: { input_tokens: 1, output_tokens: 0 },
},
});
writeSse(response, { type: 'content_block_start', index: 0, content_block: { type: 'text', text: '' } });
writeSse(response, { type: 'content_block_delta', index: 0, delta: { type: 'text_delta', text: 'PI_PROVIDER_OK' } });
finishResponse(response, [
{ type: 'content_block_stop', index: 0 },
{
type: 'message_delta',
delta: { stop_reason: 'end_turn', stop_sequence: null },
usage: { output_tokens: 1 },
},
{ type: 'message_stop' },
]);
}
function protocolFromPath(pathname) {
return MAKELore_PROVIDER_PROTOCOLS.find((protocol) => pathname.startsWith(`/${protocol}/`));
}
async function startContractServer(requests) {
const server = createServer(async (request, response) => {
response.once('error', () => undefined);
try {
const body = await readRequestBody(request);
const protocol = protocolFromPath(request.url ?? '');
if (!protocol) {
response.writeHead(404).end();
return;
}
requests.push({
protocol,
method: request.method,
path: request.url,
headers: request.headers,
model: body.model,
hasImage: JSON.stringify(body).includes(IMAGE_BASE64),
});
if (protocol === 'anthropic-messages') respondAnthropic(response, body);
else if (protocol === 'openai-responses') respondOpenAiResponses(response, body);
else respondOpenAiCompletions(response, body);
} catch (error) {
if (!response.headersSent) response.writeHead(400, { 'content-type': 'application/json' });
if (!response.destroyed && !response.writableEnded) {
response.end(JSON.stringify({ error: error instanceof Error ? error.message : String(error) }));
}
}
});
await new Promise((resolvePromise, reject) => {
server.once('error', reject);
server.listen(0, '127.0.0.1', resolvePromise);
});
const address = server.address();
if (!address || typeof address === 'string') throw new Error('Local provider server did not expose a TCP port');
return {
port: address.port,
close: () => new Promise((resolvePromise, reject) => {
server.close((error) => (error ? reject(error) : resolvePromise()));
}),
};
}
function expectedPath(protocol) {
const endpoint = protocol === 'openai-responses'
? 'responses'
: protocol === 'anthropic-messages'
? 'messages'
: 'chat/completions';
return `/${protocol}/v1/${endpoint}`;
}
function validateCapturedRequests(protocol, modelId, requests) {
const matching = requests.filter((request) => request.protocol === protocol);
const authHeader = protocol === 'anthropic-messages' ? 'x-api-key' : 'authorization';
const expectedAuth = protocol === 'anthropic-messages' ? LOCAL_API_KEY : `Bearer ${LOCAL_API_KEY}`;
const problems = [];
if (matching.length !== 4) problems.push(`expected 4 requests, got ${matching.length}`);
if (matching.some((request) => request.method !== 'POST')) problems.push('non-POST request');
if (matching.some((request) => request.path !== expectedPath(protocol))) problems.push('unexpected endpoint path');
if (matching.some((request) => request.headers[authHeader] !== expectedAuth)) problems.push('credential header mismatch');
if (matching.some((request) => request.headers[CUSTOM_HEADER] !== protocol)) problems.push('custom header mismatch');
if (matching.some((request) => request.model !== modelId)) problems.push('model mismatch');
if (matching.filter((request) => request.hasImage).length < 2) problems.push('image payload missing from initial turns');
if (problems.length > 0) throw new Error(`${protocol} local contract failed: ${problems.join('; ')}`);
return {
requestCount: matching.length,
method: 'POST',
path: expectedPath(protocol),
credentialHeader: authHeader,
customHeader: CUSTOM_HEADER,
modelId,
imageRequests: matching.filter((request) => request.hasImage).length,
};
}
export async function runLocalProviderContracts(options, projectRoot = process.cwd()) {
const scratchRoot = await mkdtemp(join(tmpdir(), 'makelore-pi-provider-contracts-'));
const requests = [];
const previousApiKey = process.env[LOCAL_API_KEY_ENV];
const server = await startContractServer(requests);
process.env[LOCAL_API_KEY_ENV] = LOCAL_API_KEY;
try {
const sourceRuntime = await resolvePiRuntime(projectRoot);
const identity = await validatePiIdentity(sourceRuntime);
const runtime = options.electronExecutablePath
? {
...sourceRuntime,
electronExecutable: options.electronExecutablePath,
cliPath: options.cliPath,
}
: sourceRuntime;
const imagePath = join(scratchRoot, 'pixel.png');
await writeFile(imagePath, Buffer.from(IMAGE_BASE64, 'base64'));
const protocols = [];
for (const protocol of MAKELore_PROVIDER_PROTOCOLS) {
const protocolRoot = join(scratchRoot, protocol);
await mkdir(protocolRoot);
const fixturePath = join(protocolRoot, 'fixture.json');
const modelId = `makelore-local-${protocol}`;
const baseUrl = protocol === 'anthropic-messages'
? `http://127.0.0.1:${server.port}/${protocol}`
: `http://127.0.0.1:${server.port}/${protocol}/v1`;
await writeFile(fixturePath, `${JSON.stringify({
id: `makelore-local-${protocol}`,
apiProtocol: protocol,
baseUrl,
apiKeyEnv: LOCAL_API_KEY_ENV,
headers: { [CUSTOM_HEADER]: protocol },
model: { id: modelId, input: ['text', 'image'] },
}, null, 2)}\n`);
const qualification = await runProviderQualification(
runtime,
protocolRoot,
fixturePath,
imagePath,
options.timeoutMs,
);
protocols.push({
protocol,
qualification,
requestContract: validateCapturedRequests(protocol, modelId, requests),
});
}
const report = {
schemaVersion: 1,
generatedAt: new Date().toISOString(),
identity,
runtime: {
artifact: options.artifactLabel ?? 'workspace-install',
executable: runtime.electronExecutable,
cliPath: runtime.cliPath,
},
scope: {
network: '127.0.0.1 only',
realProvider: false,
limitation: 'Validates Pi HTTP/SSE request contracts and worker behavior; does not qualify a real provider account.',
},
protocols,
result: 'pass',
};
if (options.reportPath) {
await mkdir(dirname(options.reportPath), { recursive: true });
await writeFile(options.reportPath, `${JSON.stringify(report, null, 2)}\n`);
}
return report;
} finally {
if (previousApiKey === undefined) delete process.env[LOCAL_API_KEY_ENV];
else process.env[LOCAL_API_KEY_ENV] = previousApiKey;
await server.close();
await rm(scratchRoot, { recursive: true, force: true, maxRetries: 5, retryDelay: 200 });
}
}
async function main() {
const options = parseProbeArgs(process.argv.slice(2));
if (options.help) {
process.stdout.write('Usage: node scripts/probe-pi-provider-contracts.mjs [probe options]\n');
return;
}
if (options.stage || options.keepStage || options.providerFixturePath || options.imagePath) {
throw new Error('Local provider contract probe owns its fixtures and does not accept staging/provider fixture flags');
}
const report = await runLocalProviderContracts(options);
process.stdout.write(`${JSON.stringify(report, null, 2)}\n`);
}
const isMain = process.argv[1]
&& pathToFileURL(resolve(process.argv[1])).href === import.meta.url;
if (isMain) {
main().catch((error) => {
process.stderr.write(`${error.stack ?? error.message}\n`);
process.exitCode = 1;
});
}

1102
scripts/probe-pi-runtime.mjs Normal file

File diff suppressed because it is too large Load Diff

View File

@@ -0,0 +1,149 @@
import { describe, expect, it } from 'vitest';
import { join, sep } from 'node:path';
import { packagedResourcesDirectory } from '../../scripts/probe-pi-packaged-runtime.mjs';
import {
buildEvidenceWaivers,
buildMissingEvidence,
buildPiProviderConfig,
createStrictJsonlParser,
lockEntrySupportsCurrentPlatform,
MAKELore_PROVIDER_PROTOCOLS,
mapMakeloreProtocolToPi,
parseProbeArgs,
percentile,
summarizeMeasurements,
} from '../../scripts/probe-pi-runtime.mjs';
describe('Pi runtime qualification probe', () => {
it('locates macOS resources beside Contents/MacOS', () => {
const executable = join('probe-root', 'MakelorePiProbe.app', 'Contents', 'MacOS', 'MakelorePiProbe');
const resources = packagedResourcesDirectory(executable, 'darwin');
expect(resources.split(sep).slice(-3)).toEqual([
'MakelorePiProbe.app',
'Contents',
'Resources',
]);
});
it('parses strict LF-delimited JSONL without splitting Unicode separators', () => {
const records: unknown[] = [];
const parser = createStrictJsonlParser((record) => records.push(record));
const source = `${JSON.stringify({ text: 'left\u2028middle\u2029right' })}\n${JSON.stringify({ ok: true })}\r\n`;
const bytes = Buffer.from(source);
parser.push(bytes.subarray(0, 9));
parser.push(bytes.subarray(9, 25));
parser.push(bytes.subarray(25));
parser.finish();
expect(records).toEqual([
{ text: 'left\u2028middle\u2029right' },
{ ok: true },
]);
});
it('uses nearest-rank p50 and p95 measurements', () => {
expect(percentile([5, 1, 4, 2, 3], 50)).toBe(3);
expect(percentile([5, 1, 4, 2, 3], 95)).toBe(5);
expect(summarizeMeasurements([1.2, 2.8, 9.1])).toEqual({
samples: 3,
p50: 3,
p95: 9,
max: 9,
});
});
it('requires explicit, bounded CLI arguments', () => {
expect(parseProbeArgs(['--stage', '--samples', '3', '--timeout-ms', '2000'])).toMatchObject({
stage: true,
samples: 3,
timeoutMs: 2000,
});
expect(() => parseProbeArgs(['--samples', '0'])).toThrow(/positive integer/);
expect(() => parseProbeArgs(['--keep-stage'])).toThrow(/requires --stage/);
expect(() => parseProbeArgs(['--electron-executable', 'electron.exe'])).toThrow(/provided together/);
expect(() => parseProbeArgs([
'--stage',
'--electron-executable',
'electron.exe',
'--cli-path',
'app.asar/dist/cli.js',
])).toThrow(/cannot be combined/);
expect(parseProbeArgs([
'--electron-executable',
'electron.exe',
'--cli-path',
'app.asar/dist/cli.js',
'--artifact-label',
'controlled-app-asar',
])).toMatchObject({ artifactLabel: 'controlled-app-asar' });
});
it('covers every current Makelore provider protocol', () => {
expect(MAKELore_PROVIDER_PROTOCOLS).toEqual([
'openai-completions',
'openai-responses',
'anthropic-messages',
'openrouter',
]);
expect(mapMakeloreProtocolToPi('openai-responses')).toEqual({
api: 'openai-responses',
compat: undefined,
});
expect(mapMakeloreProtocolToPi('openrouter')).toEqual({
api: 'openai-completions',
compat: {
thinkingFormat: 'openrouter',
sessionAffinityFormat: 'openrouter',
},
});
});
it('reports the user-approved macOS deferral separately from Phase-0 aggregation', () => {
expect(buildMissingEvidence()).toContain(
'Phase-0 artifact aggregation: Windows x64 and Linux x64; macOS x64/arm64 deferred to PI-150 by user decision on 2026-08-22',
);
expect(buildMissingEvidence()).not.toContain('Linux unpackaged and packaged samples');
});
it('reports real Provider validation as explicitly waived without claiming a real turn', () => {
expect(buildMissingEvidence()).not.toEqual(expect.arrayContaining([
expect.stringContaining('real-provider'),
expect.stringContaining('real provider'),
]));
expect(buildEvidenceWaivers()).toEqual(expect.arrayContaining([
expect.stringContaining('QG-004/QG-005'),
expect.stringContaining('explicitly waived'),
expect.stringContaining('remain unverified'),
]));
});
it('builds a Pi provider model using an environment reference, not a credential value', () => {
const config = buildPiProviderConfig({
id: 'probe-openrouter',
apiProtocol: 'openrouter',
baseUrl: 'https://openrouter.ai/api/v1',
apiKeyEnv: 'PROBE_OPENROUTER_KEY',
headers: { 'X-OpenRouter-Title': 'Makelore qualification' },
model: { id: 'anthropic/claude-sonnet-4' },
});
expect(config.providers['probe-openrouter']).toMatchObject({
api: 'openai-completions',
apiKey: '$PROBE_OPENROUTER_KEY',
models: [{
id: 'anthropic/claude-sonnet-4',
compat: {
thinkingFormat: 'openrouter',
sessionAffinityFormat: 'openrouter',
},
}],
});
});
it('rejects lock entries for a different operating system', () => {
expect(lockEntrySupportsCurrentPlatform({ os: [`!${process.platform}`] })).toBe(false);
expect(lockEntrySupportsCurrentPlatform({ os: [process.platform], cpu: [process.arch] })).toBe(true);
});
});