agilab-evidence-contracts

Capture, standardize, and replay experimental evidence with stable schemas and artifact tracking.

19|2|Updated Mar 14, 2025
One-click install
npx skills add https://github.com/ThalesGroup/agilab --skill agilab-evidence-contracts
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: agilab-evidence-contracts
Source: https://github.com/ThalesGroup/agilab/tree/main/.claude/skills/agilab-evidence-contracts
Command: npx skills add https://github.com/ThalesGroup/agilab --skill agilab-evidence-contracts

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Maintains verifiable, replayable evidence for AGILAB experiments, runs, notebooks, agent traces, and release proofs to enable auditability and reproducibility.

Core Features & Use Cases

  • Evidence surfaces: run manifests, proofs, notebook handoffs, agent evidence, supply-chain artifacts, and UI robot proofs to support end-to-end reproducibility.
  • Contract-driven workflow: ensures stable schemas, artifact tracking, redaction of secrets, and explicit verification or replay commands for each evidence type.
  • Use Case: when a notebook or app run needs to be audited or re-run with identical inputs, contracts provide a machine-readable record and a safe replay path.

Quick Start

Use the agilab-evidence-contracts skill to capture and replay run evidence with a single validation pass.

Frequently Asked Questions about agilab-evidence-contracts

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I capture and replay experimental evidence for auditability in machine learning workflows?

To capture and replay experimental evidence, you standardize run manifests, proofs, and agent traces into machine-readable contracts with stable schemas, artifact path tracking, and explicit verification commands to enable deterministic review and auditing.

What's the best way to maintain an audit trail for notebook handoffs and agent traces?

Maintaining an audit trail for notebook handoffs involves applying contract-driven workflows that track artifact paths, redact secrets, and document replay commands, ensuring end-to-end traceability and verifiable history across execution contexts.

How does evidence contract standardization work for reproducibility?

Evidence contract standardization works by enforcing stable schemas across run manifests, proofs, and release artifacts, capturing inputs and outputs to guarantee that experiments can be re-run with identical inputs for deterministic review.

Can I use evidence contracts to redact secrets in supply-chain artifacts before an audit?

Yes, you can use evidence contracts to redact secrets in supply-chain artifacts. The contract-driven workflow ensures secure handling by redacting sensitive information while preserving the artifact paths needed for verifiable replay.

When do I need reproducibility contracts for release artifacts and UI robot proofs?

You need reproducibility contracts for release artifacts and UI robot proofs when a run must be audited or re-run with identical inputs, providing a machine-readable record and a safe replay path for compliance verification.

Does evidence contract validation support deterministic replay across different execution contexts?

Yes, evidence contract validation supports deterministic replay across different execution contexts by tracking artifact paths and documenting explicit verification commands, ensuring end-to-end traceability for runs, notebooks, and agent traces.