M2: skill contracts, Python skill service, L2 eval harness with baseline
- packages/domain: ForecastRequest, BidOptimizationRequest/Result,
ReportRequest, SkillReport (+ golden and invalid fixtures, exported to
contracts/ and regenerated as pydantic models).
- skills-py/vpp_skills: FastAPI service with versioned registry; load/PV/
price forecasts (same-day-type EWM point forecast, conformal residual
quantiles — coverage test as acceptance gate); bid-optimization MILP on
HiGHS (binary block participation, hard ledger energy bounds, exact
Decimal fit of the rounded curve inside the bounds, revenue distribution
over quantile paths); report generator whose every figure is a
{tool_call_id, path} reference, with a verifier. 48 tests incl. hypothesis
property test that bids respect ledger constraints.
- packages/services: LedgerService.dayAheadBounds (the P7 cascade band
handed to the optimizer); Decimal resolved once for CJS/ESM interop.
- packages/evals: L2 metrics (MAPE, nRMSE, coverage, direction accuracy,
naive/hindsight revenue baselines), HTTP skill client, rolling-origin
harness that pushes each bid through the real ledger, CLI with
--check/--write-baseline; committed baseline on the SYNTHETIC dataset
(no historical Hubei data yet — baselines measure the harness, not KPI).
- CI: evals job boots the skill service and fails on baseline digest drift.
- docs/open-questions: A6 (flexibility marginal cost = offer floor); A4/B6
wired as placeholders. README/CLAUDE.md status → M2 done, M3 next.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01UoYoGYzHkFyv3ALenkRPhA
2026-09-02 06:29:08 -04:00
|
|
|
|
"""Python skill services (docs/05 §1–2, ADR-0001).
|
|
|
|
|
|
|
|
|
|
|
|
Every skill is typed (generated pydantic contracts in ``vpp_contracts``),
|
|
|
|
|
|
versioned (``SKILL_VERSIONS``) and stateless. No module here may import or call
|
|
|
|
|
|
an LLM: skills are deterministic numeric services and sit on the control path
|
|
|
|
|
|
(CLAUDE.md hard rule 2).
|
|
|
|
|
|
"""
|
|
|
|
|
|
|
|
|
|
|
|
SKILL_VERSIONS: dict[str, str] = {
|
|
|
|
|
|
"load-forecast": "1.0.0",
|
|
|
|
|
|
"pv-forecast": "1.0.0",
|
|
|
|
|
|
"price-forecast": "1.0.0",
|
|
|
|
|
|
"bid-optimization-milp": "1.0.0",
|
|
|
|
|
|
"report-generator": "1.0.0",
|
M4: resource agent, envelopes live, review loop, insight cards
- packages/domain: AwardNotice, DISPATCH_PLAN proposal payload, ExecutionReport,
MeteringRecord, potential-assessment and dispatch-optimization contracts,
ReviewFinding (+ typed writebacks), SemanticMemoryEntry,
EnvelopeChangeRequest, InsightCard — exported with fixtures on both sides.
- skills-py: potential-assessment (certified × rolling fulfilment, evidence
days) and dispatch-optimization (per-interval LP on HiGHS, shortfall
reported) skills + routes + tests.
- packages/services: dispatch rules in the policy pack (over-allocation,
award anchor, lineage integrity for allocations); PowerBalanceSimulator;
SimulationGateway (permit-only, idempotent, seeded execute → ExecutionReports);
envelope deviation-streak suspension + apply(); ReviewService (attribution,
reliability EWMA writeback, semantic memory, envelope recommendations as
change requests); dispatch assembler; skill client methods.
- packages/runtime: resource agent; award-decomposition, review and
envelope-review workflows; lifecycle selects simulator/gateway by proposal
type; trigger hooks for awards, execution reports, metering; decide()
resumes either lifecycle or envelope-review runs; insight cards API.
- Tests: docs/07 D-1 16:00 and D+1 end to end; reliability score 0.9 → 0.880
and the next assessment de-rates capacity; envelope suspension on a seeded
3-day streak; WIDEN request applied only by a human. 184 TS + 80 Python.
- docs/open-questions: B10 (reliability/potential parameters). README and
CLAUDE.md status → M4 done, M5 next.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01UoYoGYzHkFyv3ALenkRPhA
2026-09-02 19:31:12 -04:00
|
|
|
|
"potential-assessment": "1.0.0",
|
|
|
|
|
|
"dispatch-optimization": "1.0.0",
|
M2: skill contracts, Python skill service, L2 eval harness with baseline
- packages/domain: ForecastRequest, BidOptimizationRequest/Result,
ReportRequest, SkillReport (+ golden and invalid fixtures, exported to
contracts/ and regenerated as pydantic models).
- skills-py/vpp_skills: FastAPI service with versioned registry; load/PV/
price forecasts (same-day-type EWM point forecast, conformal residual
quantiles — coverage test as acceptance gate); bid-optimization MILP on
HiGHS (binary block participation, hard ledger energy bounds, exact
Decimal fit of the rounded curve inside the bounds, revenue distribution
over quantile paths); report generator whose every figure is a
{tool_call_id, path} reference, with a verifier. 48 tests incl. hypothesis
property test that bids respect ledger constraints.
- packages/services: LedgerService.dayAheadBounds (the P7 cascade band
handed to the optimizer); Decimal resolved once for CJS/ESM interop.
- packages/evals: L2 metrics (MAPE, nRMSE, coverage, direction accuracy,
naive/hindsight revenue baselines), HTTP skill client, rolling-origin
harness that pushes each bid through the real ledger, CLI with
--check/--write-baseline; committed baseline on the SYNTHETIC dataset
(no historical Hubei data yet — baselines measure the harness, not KPI).
- CI: evals job boots the skill service and fails on baseline digest drift.
- docs/open-questions: A6 (flexibility marginal cost = offer floor); A4/B6
wired as placeholders. README/CLAUDE.md status → M2 done, M3 next.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01UoYoGYzHkFyv3ALenkRPhA
2026-09-02 06:29:08 -04:00
|
|
|
|
}
|