feat: record verdicts in pieces and check the design against the schema

From the case 02 run cancelled at 66 minutes on 2026-10-07:

- The validator records verdicts a few Checks at a time (verdict_put) and
  they are attached on submission. It timed out five times writing them in
  one reply. Verdicts on the same code survive a run cut short; verdicts on
  changed code are dropped. Its model timeout is 300 s.
- A schema review at submission, in code: every setting a Check names is
  looked up on its Element's resource type, and an unknown resource type is
  reported. Two Revisions in that run were for Checks no code could meet.
  On the two designs that passed validation it raises 2 of 96 Checks, both
  one metric name used as a value, so it is returned once and never blocks.
- Reviews run together; the Jev coverage review stays off.

Plans record both steps and the cancelled run.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
This commit is contained in:
2026-10-08 09:39:13 +01:00
parent 5adc911a82
commit 58e778ca7a
16 changed files with 691 additions and 14 deletions

View File

@@ -22,13 +22,19 @@ from app.adk.runtime import (
)
from app.states.ledger import Ledger, LedgerRefusal, PhaseStatus
from app.tools.workspace import (
FILES,
basis_of,
file_tools,
graph_text,
graph_tools,
keep_graph,
safe_relative,
verdict_tools,
verdicts_text,
working_files,
working_verdicts,
)
from app.workflows.chain import VERDICTS_FILE, read_graph
from app.workflows.deliverables import (
DEPENDENCY_GRAPH,
DeliverableKind,
@@ -49,6 +55,7 @@ from app.workflows.target_architecture import (
ORCHESTRATOR = "orchestrator"
logger = logging.getLogger(__name__)
_TARGET = DeliverableKind.TARGET_ARCHITECTURE
_CODE = DeliverableKind.DEPLOYMENT_GUIDANCE
CLOUD_READ = "cloud_read"
_NO_ARGS: Mapping[str, Any] = {"type": "object", "properties": {}}
@@ -277,6 +284,13 @@ def specialist_tools(
is_graph = kind is DeliverableKind.TARGET_ARCHITECTURE
own_files = DEPENDENCY_GRAPH[kind].working_copy is kind
reviewed = False
is_validation = kind is DeliverableKind.VALIDATION_RESULTS
judged = ((upstream or {}).get(_CODE.value) or {}).get("attachments") or {}
# The validator's own copy of the code is the same files; either marks them.
basis = basis_of(judged or working.get(FILES) or {})
if is_validation:
working_verdicts(working, basis) # drops verdicts given on other code
graph_now = read_graph(upstream)
async def submit_deliverable(content: str, attachments: Any = None) -> ToolResult:
try:
@@ -288,6 +302,12 @@ def specialist_tools(
attachments = {**working_files(working), **attachments}
if is_graph and GRAPH_ATTACHMENT not in attachments and graph_text(working):
attachments[GRAPH_ATTACHMENT] = graph_text(working)
if (
is_validation
and VERDICTS_FILE not in attachments
and verdicts_text(working)
):
attachments[VERDICTS_FILE] = verdicts_text(working)
unsafe = [name for name in attachments if not safe_relative(name)]
if unsafe:
return _refused(
@@ -400,6 +420,11 @@ def specialist_tools(
else ()
),
*(graph_tools(working) if is_graph else ()),
*(
verdict_tools(working, basis, graph_now.check_ids if graph_now else ())
if is_validation
else ()
),
*(file_tools(working) if own_files else ()),
*(gate_cloud_read(tool, allow_cloud_read) for tool in extra),
]