Skip to content

[Bug] memos-local-plugin: config validator reports schema-declared 'reasoning' keys as unknown on every boot (same class as #2247) #2474

Open

Description

@chiefmojo

Summary

Every boot logs a false unknown config key warning for each LLM slot that sets reasoning:

config.warning message="unknown config key 'llm.reasoning' (kept as-is for forward compatibility)"
config.warning message="unknown config key 'skillEvolver.reasoning' (kept as-is …)"
config.warning message="unknown config key 'entityExtractor.reasoning' (kept as-is …)"
config.warning message="unknown config key 'l3Llm.reasoning' (kept as-is …)"

reasoning is a declared field (reasoning: Type.Optional(ReasoningSchema) in LlmSchema and SkillEvolverSchema, core/config/schema.ts), and it is applied: the value passes through to the request body. Only the warning is wrong.

This is the same class as #2247 (llm.maxTokens / llm.headers). #2248 fixed that one by adding those keys to DEFAULT_CONFIG, but reasoning was not covered, and core/config/defaults.ts on main still has no reasoning entry.

Cause

pruneUnknown(raw, DEFAULT_CONFIG, …) (core/config/index.ts) decides "unknown" by membership in the defaults object, not the schema. Any Type.Optional(...) field without a default is reported as unknown, so this will recur for every future optional key.

Why it matters

A warning that fires on every boot trains operators to ignore config.warning, and a real typo (reasonning, or a key removed from the schema) then gets lost in it.

Suggested fix

Prefer deriving the known keys from the TypeBox schema: walk Type.Object properties, recursing into nested objects, and keep FREE_FORM_CONFIG_PATHS as it is. That fixes reasoning and every future optional field at once. The lighter option, following #2248, is to add a reasoning default, but the request body must still omit it when it isn't configured.

Test idea

llm.reasoning: { enabled: false } produces no warning; llm.reasonning: {} still warns; the resolved config keeps reasoning unchanged.

Activity

  1. added
    ai:taskDispatched to AI coding agent | 已派发给 AI 编码任务
    status:in-progressSomeone or AI is working on it | 人工或 AI 正在处理
    types:bugSomething isn't working | 功能异常
    on Oct 8, 2026
  2. self-assigned this
    on Oct 8, 2026
  3. Memtensor-AI commented on Oct 8, 2026

    @Memtensor-AI
    Collaborator

    🤖 AutoDev has picked up this issue and started working on it.

    Task ID: 85e5733f87f42cdc
    Working branch: bugfix/autodev-2474-20261008202913915
    Target branch: latest dev* branch
    Workflow: opsp (analysis → coding → testing → PR)

    I will post the PR link here once done. If I need more information, I will ask in the comments.

  4. Memtensor-AI commented on Oct 8, 2026

    @Memtensor-AI
    Collaborator

    ✅ AutoDev task 85e5733f87f42cdc completed.

    Summary: Fixes #2474 — the memos-local-plugin config validator falsely reported schema-declared reasoning fields (llm.reasoning, skillEvolver.reasoning, l3Llm.reasoning) as unknown config keys on every boot. The root cause was pruneUnknown in apps/memos-local-plugin/core/config/index.ts using DEFAULT_CONFIG as the "known keys" oracle, which misses any Type.Optional(...) field without a concrete default — the same defect class as #2247/#2248.

    The fix rewrites pruneUnknown to walk the TypeBox ConfigSchema directly: it uses the properties map of type: "object" nodes to decide what is known, and treats patternProperties (TypeBox Type.Record) as a free-form map so llm.headers, logging.channels, and similar free-form sections continue to accept arbitrary child keys. The existing FREE_FORM_CONFIG_PATHS allowlist is preserved for backward compatibility. The warning text, pass-through semantics, and the downstream pipeline (deepMerge → Value.Default → Value.Errors) are all unchanged.

    Verification: a new tests/unit/config/schema-driven-unknown-keys.test.ts adds 9 regression cases (reasoning no-warning across all three LLM slots, typos still warn, Record-based free-form maps unchanged, nested unknowns inside a known Optional block still warn, resolved config retains the reasoning block byte-for-byte). Full plugin unit suite 1599 pass + 1 skipped, integration suite 6/6, tsc -p tsconfig.json --noEmit clean.

    Base branch: main
    Branch: bugfix/autodev-2474-20261008202913915
    Commit: 42427d5
    PR: #2475
    Assigned to: @syzsunshine219
    Reviewers: @whipser030, @hijzy

  5. added
    ai:testingAI agent is running tests | AI 正在运行测试
    ai:failedAI task failed | AI 任务失败
    and removed
    ai:codingAI agent is coding | AI 正在编码
    ai:testingAI agent is running tests | AI 正在运行测试
    ai:taskDispatched to AI coding agent | 已派发给 AI 编码任务
    on Oct 8, 2026
  6. github-actions commented on Nov 10, 2026

    @github-actions
    Contributor

    This issue has been automatically marked as stale due to inactivity.

  7. added
    status:staleAuto-marked stale after 30d inactivity; closes 7d later unless exempt.
    on Nov 10, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Labels

ai:failedAI task failed | AI 任务失败area:pluginOpenClaw & Hermesstatus:in-progressSomeone or AI is working on it | 人工或 AI 正在处理status:staleAuto-marked stale after 30d inactivity; closes 7d later unless exempt.types:bugSomething isn't working | 功能异常

Type

No type

Projects

No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions