Add a holdout-safe Fit experience index

[divo12]

Summary

• write one content-bound, typed experience index per Fit campaign
• preserve paired developer runs, verifier feedback, trace/snapshot lineage, candidate manifest, prediction attribution, diff and raw developer benchmark references
• expose only selection/admission completion and pass decisions; never copy holdout identifiers, snapshots, outputs, feedback, or artifact paths
• validate canonical source paths, digests, and rebuilt developer evidence on cached Fit reads

Verification

• 167 tests passed
• Ruff clean
• mypy --strict clean
• Bandit clean
• pip-audit: no known vulnerabilities
• wheel and sdist build
• forbidden dynamic/raw type scan clean
• independent reviewer: APPROVE

[cubic-dev-ai]

1 issue found across 6 files

Prompt for AI agents (unresolved issues)

Check if these issues are valid — if so, understand the root cause of each and fix them. If appropriate, use sub-agents to investigate and fix each issue separately.


<file name="src/ofw/benchmarking.py">

<violation number="1" location="src/ofw/benchmarking.py:82">
P2: When a repository already contains a baseline or Fit manifest from before this change, cached reads reject its attempts because `source_case_id` is missing, and legacy baseline verification reports semantic drift after the field is added to `_semantic`. Add backward-compatible deserialization and preserve the legacy digest for old artifacts, or explicitly version and migrate/invalidate these persisted results.</violation>
</file>

<sub>You're on the cubic free plan with 18 free PR reviews remaining this month. Upgrade for unlimited reviews.

Re-trigger cubic</sub>

[cubic-dev-ai]

P2: When a repository already contains a baseline or Fit manifest from before this change, cached reads reject its attempts because source_case_id is missing, and legacy baseline verification reports semantic drift after the field is added to _semantic. Add backward-compatible deserialization and preserve the legacy digest for old artifacts, or explicitly version and migrate/invalidate these persisted results.

Prompt for AI agents

Check if this issue is valid — if so, understand the root cause and fix it. At src/ofw/benchmarking.py, line 82:

<comment>When a repository already contains a baseline or Fit manifest from before this change, cached reads reject its attempts because `source_case_id` is missing, and legacy baseline verification reports semantic drift after the field is added to `_semantic`. Add backward-compatible deserialization and preserve the legacy digest for old artifacts, or explicitly version and migrate/invalidate these persisted results.</comment>

<file context>
@@ -79,6 +79,7 @@ def digest(self) -> Sha256Digest:
 @dataclass(frozen=True, slots=True)
 class CaseAttempt:
     case_id: str
+    source_case_id: str
     partition: ExportPartition
     critical: bool
</file context>