fix(skills): improve adapter normalization in data2dsl_skill [ticket-027]

[MatthiasLew]

Summary

• Update _normalize_raw\ in \src/data2dsl_skill.py\ to route raw inputs to typed adapters.
• Wrap contract self-test execution in clean PASS/FAIL response.
• Index ticket-027 in project/TICKETS.md.

Validation

• \project\governance-check.bat\ passed (GOV-PASS).
• \pytest tests/\ passed (15/15 passed).

Signed-off-by: agent:antigravity

[ifuri-validator-agent]

Deterministic Validator approval for exact head ac1f1fa054b93f006ca2bbe36c8d14e386e45cb8.

Ticket: ticket-027
Correlation ID: data2dsl-pr-28-ticket-027-ac1f1fa0
Model: openrouter/z-ai/glm-5.2
Reviewed diff chunks: 2
Advisory LLM verdict: REQUEST_CHANGES
Advisory summary: Reviewed all 2 diff chunk(s). Diff adds a _normalize_raw helper that dispatches to concrete adapters and refactors self_test to catch exceptions; ticket metadata is added. The change looks reasonable, but the central execute_compare routing that consumes _normalize_raw is truncated/incomplete in the visible chunk, so I cannot fully verify correctness of the adapter invocations and downstream usage. | This chunk updates execute_compare to use the new _normalize_raw helper and DeterministicComparator, returning both result and bundle. The changes are internally consistent with the PR intent. No security issues, no secrets, no workflow modifications, and no failing tests present.
Advisory findings: Cannot fully validate _normalize_raw integration: the execute_compare method where _normalize_raw is used is only partially visible (ends mid-method at the left_observation resolution), so the correctness of how the normalized result flows into the comparator cannot be assessed.; The old import of normalize_observation and compare was replaced with DeterministicComparator plus several adapter imports, but the visible chunk does not confirm the comparator is instantiated/invoked consistently with the new adapter-based approach, risking a runtime mismatch or NameError if compare() was previously used elsewhere in the unseen portion.; The markdown branch assumes query['subject']['actor'] exists; KeyError propagation in a helper that is not caught at the self_test layer (which only wraps contract_self_test) may surface as an unhandled crash if a caller passes a malformed query.
The LLM output above is advisory and was not used as the approval trust root.
Merge will be attempted after this approval when explicitly authorized.

Decision record (recomputable)

DECISION D-027-9712
TICKET ticket-027
HEAD_SHA ac1f1fa054b93f006ca2bbe36c8d14e386e45cb8
CORRELATION_ID data2dsl-pr-28-ticket-027-ac1f1fa0
ACTOR agent:ifuri-validator-agent[bot]
APPLIED_RULE P-CORE-015
INPUT author_login = "MatthiasLew"
INPUT observed_checks = ["governance / remote lifecycle=PASS"]
INPUT required_checks = ["governance / remote lifecycle"]
INPUT required_checks_source = "protected registry + GitHub applied rules (env/request)"
INPUT reviewer_login = "ifuri-validator-agent[bot]"
VERDICT APPROVE AUTHORITY DETERMINISTIC
REJECTED REQUEST_CHANGES BECAUSE NO_UNSAFE_CHANGE_REASON_FOUND
ADVISORY llm_verdict = "REQUEST_CHANGES" MODEL "openrouter/z-ai/glm-5.2"
ASSERT VERDICT_AUTHORITY != "ADVISORY"