Agent Skills by moffran
Showing 50 vetted skills indexed across 1 GitHub repositories.
Moffrance-standards-gap-analyzer
Verify STD decisions are implemented in code and documentation.
Moffrance-regulatory-compliance
Map calibrated_explanations capabilities to EU regulatory obligations for audit-ready compliance documentation.
Moffrance-release-planner
Generate structured vX.Y.Z implementation plans from RELEASE_PLAN_v1.md with ADR and STD references.
Moffrance-fallback-test
Verify fallback visibility with pytest.warns and caplog in pytest tests.
Moffrance-adr-consult
Identify applicable ADRs and extract binding constraints from docs/improvement/adrs/.
Moffrance-notebook-audit
Scan notebooks for private-member usage, policy violations, and rendering errors.
Moffrance-serializer-impl
Serialize calibrator state with ADR-031 primitives and sha256 checksums.
Moffrance-classification
Calibrate binary and multiclass classification predictions with Venn-Abers calibration.
Moffrance-onboard
Guide new CE users through core invariants and session routing at startup.
Moffrance-release-task
Identify, implement, and verify the next release task from plan files.
Moffrance-test-creator
Generate targeted tests for documented coverage gaps using the ADR-030 workflow.
Moffrance-data-preparation
Validate and preprocess mixed-type data with categorical encoding and mapping export.
Moffrance-adr-gap-analyzer
Identify gaps between ADR intent and realized implementation and documentation.
Moffrance-skill-registry-sync
Synchronize skill registries with the filesystem inventory in CI workflows.
Moffrance-devils-advocate
Critique code-change proposals for hidden risks and regressions with structured ratings.
Moffrance-rtd-auditor
Audits ReadTheDocs projects for navigation, accuracy, and ADR consistency, generating severity-ranked findings.
Moffrance-test-pruning-expert
Identify redundant tests using unique-line counts and overlap fingerprints.
Moffrance-mondrian-conditional
Configure and validate Mondrian conditional calibration for subgroup-specific uncertainty estimates.
Moffrance-reject-policy
Integrate configurable reject and defer policies into ML prediction workflows.
Moffrance-factual-explain
Generate factual Calibrated Explanations for model predictions with explain_factual workflows.
Moffrance-plot-review
Audit plotting code against CE Plot Review ADR dimensions and generate conformance findings.
Moffrance-payload-governance
Validate explanation payload schemas and provenance against ADR-005 and ADR-028 invariants.
Moffrance-test-author
Standardize CE test authoring with AAA structure and public-API focus.
Moffrance-rtd-writer
Create and maintain ReadTheDocs pages with CE-first structure and toctree integration.