[PR #1624] [MERGED] test(clients): verify Copilot compact schema workflows #4709

New issue

Closed

opened 2026-10-31 00:54:13 +00:00 by bsakaguchi · 0 comments

bsakaguchi commented

2026-10-31 00:54:13 +00:00

Owner

📋 Pull Request Information

Original PR: https://github.com/morluto/rea/pull/1624
Author: @N0zoM1z0
Created: 10/10/2026
Status: ✅ Merged
Merged: 10/10/2026
Merged by: @N0zoM1z0

Base: main ← Head: fix/copilot-static-catalog-20261010

📝 Commits (1)

d39d161 test(clients): verify Copilot compact schema workflows

📊 Changes

3 files changed (+46 additions, -4 deletions)

View changed files

📝 docs/installation.md (+12 -0)
📝 docs/testing.md (+14 -3)
📝 scripts/verify-copilot-client.mjs (+20 -1)

📄 Description

Copilot CLI 1.0.95 with GPT-4.1 metadata still blocks REA's full input-schema catalog before the first model request. Its existing compact-profile alternative now has a documented configuration recipe and repeatable native-client verification for both ordinary chat and analysis.

The optional verifier accepts REA_VERIFY_COPILOT_SCHEMA_PROFILE=compact, applies it only to its disposable MCP registration after the setup/idempotence checks, and records the selected profile. The full profile remains the default. All canonical tools are forwarded and every model-facing input schema is validated. Compact advertisements can reduce nested constraints; canonical server-side validation remains complete. This is an alternative workflow, not resolution of the full-profile boundary in #1554, which stays open. No production contracts or default CI lanes change.

Validation on Linux x64, Node 24.18.0, native @github/copilot@1.0.95, current 170-tool catalog:

GPT-4.1 / compact: native ordinary chat and skill-driven JavaScript MCP analysis passed, including Unicode target paths and complete Evidence retrieved through Copilot's native spill reader.
GPT-5.4 / full: native analysis and spill retrieval passed as a default-profile regression.
All three successful runs verified empty owned-process lineages; receipts retain the selected model/profile and captured requests.
GPT-4.1 / full reproduced compaction_static_context_blocked before HTTP.
npm run test:focused -- tests/boundary/mcp/compactInputSchemaMcp.test.ts tests/boundary/mcp/toolSchemaValidity.test.ts — 15 passed.
npm run check:fast passed.

Inference uses a deterministic loopback OpenAI-compatible fixture. Live provider APIs, native Windows and default OS-home skill discovery are unverified. No real Hopper/Ghidra verification was performed; the actual JavaScript provider supplies analysis.

Related: #1554.

_{🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.}

## 📋 Pull Request Information **Original PR:** https://github.com/morluto/rea/pull/1624 **Author:** [@N0zoM1z0](https://github.com/N0zoM1z0) **Created:** 10/10/2026 **Status:** ✅ Merged **Merged:** 10/10/2026 **Merged by:** [@N0zoM1z0](https://github.com/N0zoM1z0) **Base:** `main` ← **Head:** `fix/copilot-static-catalog-20261010` --- ### 📝 Commits (1) - [`d39d161`](https://github.com/morluto/rea/commit/d39d161744d7713ec32e9ba61d1b7c58df4c94df) test(clients): verify Copilot compact schema workflows ### 📊 Changes **3 files changed** (+46 additions, -4 deletions) <details> <summary>View changed files</summary> 📝 `docs/installation.md` (+12 -0) 📝 `docs/testing.md` (+14 -3) 📝 `scripts/verify-copilot-client.mjs` (+20 -1) </details> ### 📄 Description Copilot CLI 1.0.95 with GPT-4.1 metadata still blocks REA's full input-schema catalog before the first model request. Its existing compact-profile alternative now has a documented configuration recipe and repeatable native-client verification for both ordinary chat and analysis. The optional verifier accepts `REA_VERIFY_COPILOT_SCHEMA_PROFILE=compact`, applies it only to its disposable MCP registration after the setup/idempotence checks, and records the selected profile. The full profile remains the default. All canonical tools are forwarded and every model-facing input schema is validated. Compact advertisements can reduce nested constraints; canonical server-side validation remains complete. This is an alternative workflow, not resolution of the full-profile boundary in #1554, which stays open. No production contracts or default CI lanes change. Validation on Linux x64, Node 24.18.0, native `@github/copilot@1.0.95`, current 170-tool catalog: - GPT-4.1 / compact: native ordinary chat and skill-driven JavaScript MCP analysis passed, including Unicode target paths and complete Evidence retrieved through Copilot's native spill reader. - GPT-5.4 / full: native analysis and spill retrieval passed as a default-profile regression. - All three successful runs verified empty owned-process lineages; receipts retain the selected model/profile and captured requests. - GPT-4.1 / full reproduced `compaction_static_context_blocked` before HTTP. - `npm run test:focused -- tests/boundary/mcp/compactInputSchemaMcp.test.ts tests/boundary/mcp/toolSchemaValidity.test.ts` — 15 passed. - `npm run check:fast` passed. Inference uses a deterministic loopback OpenAI-compatible fixture. Live provider APIs, native Windows and default OS-home skill discovery are unverified. No real Hopper/Ghidra verification was performed; the actual JavaScript provider supplies analysis. Related: #1554. --- <sub>🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.</sub>

bsakaguchi

2026-10-31 00:54:13 +00:00

closed this issue
added the
pull-request
label

bsakaguchi referenced this issue

2026-10-31 01:13:58 +00:00

[PR #4721] [CLOSED] fix(electron): exclude DOM nodes with out-of-range parentIndex #6372

No milestone

No project

No assignees

1 participant

Notifications

Due date

The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference

morluto/rea#4709

No description provided.

Rows
Columns