fix(llm): let the task schema disambiguate JSON candidates in model prose

[zzwong]

Closes #608

Problem

Structured-output decoding already recovers a JSON object from surrounding prose, via extractSingleJSONObject. It requires exactly one syntactically valid top-level object, deliberately, so ambiguous output is never guessed at.

A reasoning preamble routinely breaks that. Running the resident extractor against representative inputs:

preamble before the answer	recovered today
no braces here	yes
draft: {"decision":"skip"}	no — two candidates
I must return JSON, not {}.	no — {} is a candidate
answer itself truncated	no — correctly unrecoverable

The budget is two validation attempts and an exhausted task is blocking, so a model that drafts before answering loses half its retry budget systematically rather than occasionally.

Change

Extraction returns every balanced valid candidate, and the task's own decoder picks among them: exactly one passing the schema is accepted, zero or several keep the existing error. A lone failing candidate still reports its own schema error, since that describes the real violation.

This cannot loosen validation. A candidate failing DisallowUnknownFields, or carrying the wrong thread_id or schema_version, was never a possible answer — only the bytes reaching the decoder change, never what the decoder accepts. Trailing-data rejection is unaffected: each candidate is a single balanced object, and the raw-text check still runs first.

Evidence

Every behavior change mutation-checked by reverting it alone:

Mutation	Fails
Restore the single-candidate rule	recovers_the_schema-valid_object_when_the_preamble_drafts_another
Accept when passed >= 1	rejects_two_schema-valid_objects
Skip the i = end advance	TestExtractJSONObjects/nested_objects_and_arrays
Ignore json.Valid	prose_braces_alongside_one_valid_object, valid_object_nested_in_invalid_prose_braces

go build ./..., go vet ./internal/... and go test -count=1 ./... are clean.

Note on scope

The originally filed issue claimed no extraction existed; that was wrong and the issue has been corrected. A truncated response remains unrecoverable by extraction and is the retry path's job, not this one's.

[monit-reviewer]

Automated PR Review

Reviewed commit: 96512a381c81
Profile: open-cli-collective-review - Posting as: monit-reviewer

Summary

Reviewer	Findings
go:implementation-tests	0

Reviewer Coverage

Complete coverage details are retained in the local coverage.json artifact.

• go:implementation-tests — complete (broad); scope: 4 files; examples: internal/llm/adapter.go, internal/llm/adapter_test.go, internal/llm/extract.go, internal/llm/extract_test.go; body-inspected: 4 files; examples: internal/llm/adapter.go, internal/llm/adapter_test.go, internal/llm/extract.go, internal/llm/extract_test.go; skipped: none; context (not coverage): 1 file; examples: internal/llm/contracts.go; constraints: Reviewed pinned source and tests only; build, vet, and tests were not executed with the available read-only tools.

0 PR discussion threads considered. 0 summarized; 0 resolved.

---

Completed in 34s | $0.03 | openai-codex/gpt-6.1-sol | cr 0.10.328

Field	Value
Model	openai-codex/gpt-6.1-sol
Reviewers	go:implementation-tests
Engine	pi_rpc · openai-codex/gpt-6.1-sol
Reviewed by	cr · monit-reviewer
Duration	34s wall · 24s compute
Cost	$0.03
Tokens	12.0k in / 286 out

Per-workstream usage

• orchestrator-selection — openai-codex/gpt-6.1-sol
  • In: 5.6k
  • Out: 133
  • Cache read: 0
  • Cache create: 0
  • Cost: $0.01
  • Duration: 5s
• go:implementation-tests — openai-codex/gpt-6.1-sol
  • In: 2.7k
  • Out: 103
  • Cache read: 22.9k
  • Cache create: 0
  • Cost: $0.01
  • Duration: 16s
• orchestrator-rollup — openai-codex/gpt-6.1-sol
  • In: 3.7k
  • Out: 50
  • Cache read: 0
  • Cache create: 0
  • Cost: $0.01
  • Duration: 3s