generated from coulomb/repo-seed
Prefer JSON-bearing envelope fields, skip metadata, in Claude CLI unwrap
The first CUST-WP-0045 canary retry after 9de0f49 still failed schema
validation with `Expecting value: line 1 column 1 (char 0)`. The original
allowlist returned envelope.result verbatim, which on longer prompts
carries the model's conversational preamble ("Triage report generated
and returned via structured output. Key signals: ..."), not the
schema-enforced JSON. The actual structured payload lives in a different
envelope field whose name varies across CLI versions.
Make the unwrap order-aware:
1. Scan envelope fields and return the first one whose value parses as
JSON (dict, list, or a string that loads cleanly). Skip well-known
metadata keys (type, usage, total_cost_usd, etc.) so telemetry can
never be mistaken for the model payload.
2. Fall back to the original text-field allowlist only when no field
carries JSON, so non-schema callers via this same code path still
see the model's prose.
3. Surface the raw envelope as last resort.
This is robust against unknown envelope shapes — as long as the schema-
enforced JSON appears somewhere in a non-metadata field, the adapter
will find it.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
This commit is contained in:
@@ -192,20 +192,33 @@ def _json_schema_arg(config: RunConfig) -> str | None:
|
||||
return None
|
||||
|
||||
|
||||
# Field names Claude Code's `--output-format json` envelope is known to use
|
||||
# for the model's primary textual response. Probed in order; the first match
|
||||
# wins. If none match (because the envelope shape is something we haven't
|
||||
# seen), we return the raw envelope string so the caller still gets the data
|
||||
# and can introspect it.
|
||||
# Envelope field names Claude Code's `--output-format json` is known to use
|
||||
# for the model's primary textual response. Used as a fall-back when no field
|
||||
# carries a JSON-parseable payload (e.g. plain prose generation).
|
||||
_ENVELOPE_TEXT_FIELDS = ("result", "result_text", "content", "text", "output")
|
||||
|
||||
|
||||
def _unwrap_cli_json_envelope(stdout: str, config: RunConfig) -> str:
|
||||
"""Extract the model's payload from Claude CLI's --output-format json envelope.
|
||||
|
||||
Only attempts unwrap when --json-schema was set, because that's the only
|
||||
code path that adds --output-format json to the CLI invocation. Other
|
||||
paths keep raw stdout (current behavior preserved).
|
||||
Only runs when --json-schema was set (the only code path that adds
|
||||
--output-format json to the CLI invocation). Other callers keep the raw
|
||||
stdout behavior unchanged.
|
||||
|
||||
Strategy: when --json-schema is set the caller wants JSON back, so prefer
|
||||
any envelope field whose value is itself valid JSON (dict, list, or a
|
||||
string that parses as JSON). This handles two observed envelope shapes:
|
||||
|
||||
1. Short prompts where the model emits the structured payload directly
|
||||
in the `result` field as a JSON-encoded string.
|
||||
2. Longer prompts where the model emits a conversational preamble in
|
||||
`result` and the schema-enforced JSON in a separate field (the exact
|
||||
field name varies across CLI versions).
|
||||
|
||||
Fall back to the first text field only when no JSON-bearing field exists,
|
||||
so non-schema callers via this code path still see the model's prose.
|
||||
Surface the raw envelope as a last resort so the operator can see what
|
||||
shape arrived and extend the strategy.
|
||||
"""
|
||||
if not _json_schema_arg(config):
|
||||
return stdout
|
||||
@@ -218,13 +231,47 @@ def _unwrap_cli_json_envelope(stdout: str, config: RunConfig) -> str:
|
||||
return stdout
|
||||
if not isinstance(envelope, dict):
|
||||
return stdout
|
||||
|
||||
json_payload = _find_json_payload(envelope)
|
||||
if json_payload is not None:
|
||||
return json_payload
|
||||
|
||||
for key in _ENVELOPE_TEXT_FIELDS:
|
||||
if key in envelope:
|
||||
value = envelope[key]
|
||||
if isinstance(value, str):
|
||||
return value
|
||||
if isinstance(value, (dict, list)):
|
||||
return json.dumps(value)
|
||||
# Unknown envelope shape — surface it raw so the operator can see it
|
||||
# in the validation error and we can update _ENVELOPE_TEXT_FIELDS.
|
||||
value = envelope.get(key)
|
||||
if isinstance(value, str):
|
||||
return value
|
||||
if isinstance(value, (dict, list)):
|
||||
return json.dumps(value)
|
||||
|
||||
return stdout
|
||||
|
||||
|
||||
def _find_json_payload(envelope: dict) -> str | None:
|
||||
"""Return the first envelope value that represents valid JSON.
|
||||
|
||||
Insertion order is preserved by Python dicts, so this prefers fields the
|
||||
CLI lists earliest in its envelope. Skips obvious metadata keys (cost,
|
||||
usage, timing) so we never accidentally pick a numeric or telemetry value.
|
||||
"""
|
||||
for key, value in envelope.items():
|
||||
if key in _ENVELOPE_METADATA_KEYS:
|
||||
continue
|
||||
if isinstance(value, (dict, list)):
|
||||
return json.dumps(value)
|
||||
if isinstance(value, str):
|
||||
stripped = value.strip()
|
||||
if stripped.startswith(("{", "[")):
|
||||
try:
|
||||
json.loads(stripped)
|
||||
except json.JSONDecodeError:
|
||||
continue
|
||||
return stripped
|
||||
return None
|
||||
|
||||
|
||||
# Envelope keys that carry telemetry, never the model payload.
|
||||
_ENVELOPE_METADATA_KEYS = frozenset({
|
||||
"type", "subtype", "model", "usage", "total_cost_usd", "cost_usd",
|
||||
"duration_ms", "duration_api_ms", "num_turns", "session_id",
|
||||
"is_error", "stop_reason", "permission_denials", "uuid",
|
||||
})
|
||||
|
||||
@@ -16,11 +16,13 @@ def test_execute_prompt_passes_json_schema_to_claude_cli(monkeypatch):
|
||||
calls["capture_output"] = capture_output
|
||||
calls["text"] = text
|
||||
calls["timeout"] = timeout
|
||||
return SimpleNamespace(
|
||||
returncode=0,
|
||||
stdout='{"summary":"ok","recommendations":[]}',
|
||||
stderr="",
|
||||
)
|
||||
# With --output-format json the CLI returns an envelope.
|
||||
envelope = {
|
||||
"type": "result",
|
||||
"result": '{"summary":"ok","recommendations":[]}',
|
||||
}
|
||||
import json as _json
|
||||
return SimpleNamespace(returncode=0, stdout=_json.dumps(envelope), stderr="")
|
||||
|
||||
monkeypatch.setattr("llm_connect.claude_code.subprocess.run", fake_run)
|
||||
adapter = ClaudeCodeAdapter(cli_path="/custom/claude")
|
||||
@@ -33,22 +35,18 @@ def test_execute_prompt_passes_json_schema_to_claude_cli(monkeypatch):
|
||||
),
|
||||
)
|
||||
|
||||
assert calls == {
|
||||
"cmd": [
|
||||
"/custom/claude",
|
||||
"--print",
|
||||
"--json-schema",
|
||||
'{"type":"object"}',
|
||||
"--output-format",
|
||||
"json",
|
||||
],
|
||||
"input": "Produce a report.",
|
||||
"capture_output": True,
|
||||
"text": True,
|
||||
"timeout": 42,
|
||||
}
|
||||
# Stdout shape that does not match any known envelope field is returned
|
||||
# verbatim so the caller can introspect and we can extend the field list.
|
||||
assert calls["cmd"] == [
|
||||
"/custom/claude",
|
||||
"--print",
|
||||
"--json-schema",
|
||||
'{"type":"object"}',
|
||||
"--output-format",
|
||||
"json",
|
||||
]
|
||||
assert calls["input"] == "Produce a report."
|
||||
assert calls["timeout"] == 42
|
||||
# Envelope's result field carries the schema-enforced JSON; the adapter
|
||||
# unwraps it before returning to the caller.
|
||||
assert response.content == '{"summary":"ok","recommendations":[]}'
|
||||
|
||||
|
||||
@@ -80,6 +78,53 @@ def test_execute_prompt_unwraps_cli_json_envelope_result_field(monkeypatch):
|
||||
assert response.content == '{"summary":"ok","recommendations":[]}'
|
||||
|
||||
|
||||
def test_execute_prompt_prefers_json_field_over_prose_preamble(monkeypatch):
|
||||
"""When the model adds a prose preamble in the envelope's primary text
|
||||
field but the schema-enforced JSON is in a different field, the adapter
|
||||
must find and return the JSON, not the preamble."""
|
||||
def fake_run(cmd, input, capture_output, text, timeout): # noqa: ANN001
|
||||
envelope = {
|
||||
"type": "result",
|
||||
"result": "Triage report generated and returned via structured output. Key signals: healthy.",
|
||||
"structured_result": '{"summary":"healthy","recommendations":[]}',
|
||||
"total_cost_usd": 0.002,
|
||||
}
|
||||
import json as _json
|
||||
return SimpleNamespace(returncode=0, stdout=_json.dumps(envelope), stderr="")
|
||||
|
||||
monkeypatch.setattr("llm_connect.claude_code.subprocess.run", fake_run)
|
||||
adapter = ClaudeCodeAdapter(cli_path="/custom/claude")
|
||||
|
||||
response = adapter.execute_prompt(
|
||||
"Long triage prompt.",
|
||||
RunConfig(model_params={"json_schema": {"type": "object"}}),
|
||||
)
|
||||
|
||||
assert response.content == '{"summary":"healthy","recommendations":[]}'
|
||||
|
||||
|
||||
def test_execute_prompt_skips_envelope_metadata_keys(monkeypatch):
|
||||
"""Metadata keys like `type`, `model`, `usage` must never be returned as
|
||||
the model payload, even if their values look JSON-like."""
|
||||
def fake_run(cmd, input, capture_output, text, timeout): # noqa: ANN001
|
||||
envelope = {
|
||||
"type": '{"this":"is_metadata"}', # decoy
|
||||
"usage": {"input_tokens": 5}, # decoy dict
|
||||
"result": '{"summary":"ok"}',
|
||||
}
|
||||
import json as _json
|
||||
return SimpleNamespace(returncode=0, stdout=_json.dumps(envelope), stderr="")
|
||||
|
||||
monkeypatch.setattr("llm_connect.claude_code.subprocess.run", fake_run)
|
||||
adapter = ClaudeCodeAdapter(cli_path="/custom/claude")
|
||||
|
||||
response = adapter.execute_prompt(
|
||||
"Prompt.", RunConfig(model_params={"json_schema": {"type": "object"}})
|
||||
)
|
||||
|
||||
assert response.content == '{"summary":"ok"}'
|
||||
|
||||
|
||||
def test_execute_prompt_no_unwrap_without_json_schema(monkeypatch):
|
||||
"""Without --json-schema we do not pass --output-format json, so the
|
||||
envelope unwrap path stays inert and raw stdout passes through."""
|
||||
|
||||
Reference in New Issue
Block a user