Verification, testing & evaluation
Evaluate software and agent candidates with controlled harnesses, quality plans, graders, replay artifacts and explicit release evidence.
Deterministic and live-adapter evaluation surfaces are source-backed. Simulated results, fixtures and successful retries do not establish production qualification or authorize effects.
CONNECTED ACROSS LSSM
From intent to governed execution.
Select an entry to explore its neighbours. Move between ownership, specifications, areas, products, data and infrastructure without leaving this page.
16 entries · 74 recorded relationships
Arrow keys navigate entries. Enter selects. Alt + arrows rearranges. F fits the view; Shift + F fits the selection. + and − zoom. Shift + F10 opens actions.
Accessible entry list (16)
- Products & experiences
- Products & experiences
- Shared foundations
- Shared foundations
- Inspected sources
- Toolchain & agents
- Toolchain & agents
- Toolchain & agents
- Toolchain & agents
- Toolchain & agents
- Toolchain & agents
- Toolchain & agents
- Infrastructure services
- Specialized specifications
- Specialized specifications
- Specialized specifications
Capabilities
- Specifications & delivery ↗Supports this work
An editorial capability connection. Exact implementation scope and host requirements are stated in the role boundary.
- Product management ↗Supports this work
An editorial capability connection. Exact implementation scope and host requirements are stated in the role boundary.
- Product & company metrics ↗Supports this work
An editorial capability connection. Exact implementation scope and host requirements are stated in the role boundary.
Problem areas
- Adaptable software ↗Useful in this area
Problem-based navigation for an explicitly scoped role.
- Data, research & intelligence ↗Useful in this area
Problem-based navigation for an explicitly scoped role.
Related services and proposed uses
- ContractSpec Studio ↗Related product work
Relevant product context. No deployed integration, data-sharing grant or fully autonomous execution chain is asserted.
- CompanyOS ↗Related product work
Relevant product context. No deployed integration, data-sharing grant or fully autonomous execution chain is asserted.
- App submission readiness ↗Related reusable role
Editorial navigation between complementary responsibilities. This link does not establish an installed adapter, execution permit or production binding.
- Content, media & editorial workflows ↗Related reusable role
Editorial navigation between complementary responsibilities. This link does not establish an installed adapter, execution permit or production binding.
- Builder & product delivery ↗Adjacent lifecycle role
Editorial lifecycle navigation; each handoff, provider and execution authority requires its own qualification.
- Agents & coordinated execution ↗Adjacent lifecycle role
Editorial lifecycle navigation; each handoff, provider and execution authority requires its own qualification.
- CI, release & progressive delivery ↗Adjacent lifecycle role
Editorial lifecycle navigation; each handoff, provider and execution authority requires its own qualification.
- Governed improvement & self-healing ↗Adjacent lifecycle role
Editorial lifecycle navigation; each handoff, provider and execution authority requires its own qualification.
- Policy, review & governance ↗Governed toolchain work
Shared governance context. This graph link is not an execution permit or evidence of a live binding.
Shared foundations & holding scope
- ContractSpec ↗Shared ecosystem foundation
Declared shared foundation scope. Source ownership is shown separately; this is not a measured runtime import.
- LSSM OS ↗Shared ecosystem foundation
Declared shared foundation scope. Source ownership is shown separately; this is not a measured runtime import.
Specialized specifications
- Provider routing specs ↗Source-mapped contract scope
Pinned package purpose and reviewed source map associate this contract surface with the target. This relationship is not a manifest dependency or proof of product runtime adoption.
- Infrastructure operations specs ↗Source-mapped contract scope
Pinned package purpose and reviewed source map associate this contract surface with the target. This relationship is not a manifest dependency or proof of product runtime adoption.
- LSSM OS specs ↗Source-mapped contract scope
Pinned package purpose and reviewed source map associate this contract surface with the target. This relationship is not a manifest dependency or proof of product runtime adoption.
Within LSSM
- LSSM ↗Owned by LSSM
Founder-confirmed ecosystem ownership. This records the stated organizational structure, not a company-register verification or ownership of an external provider.
Groups using this responsibility
- Company & owner assistance ↗Service context for this group
Derived through explicit member products, selected capabilities or their recorded source paths. The source edge chain preserves implemented, proposed and editorial scopes; the group link adds navigation, not runtime integration.
- Company decisions & learning ↗Service context for this group
Derived through explicit member products, selected capabilities or their recorded source paths. The source edge chain preserves implemented, proposed and editorial scopes; the group link adds navigation, not runtime integration.
- Contracts & application foundations ↗Service context for this group
Derived through explicit member products, selected capabilities or their recorded source paths. The source edge chain preserves implemented, proposed and editorial scopes; the group link adds navigation, not runtime integration.
- Product discovery & decisions ↗Service context for this group
Derived through explicit member products, selected capabilities or their recorded source paths. The source edge chain preserves implemented, proposed and editorial scopes; the group link adds navigation, not runtime integration.
- Planning & delivery ↗Service context for this group
Derived through explicit member products, selected capabilities or their recorded source paths. The source edge chain preserves implemented, proposed and editorial scopes; the group link adds navigation, not runtime integration.
- Feedback & outcome checks ↗Service context for this group
Derived through explicit member products, selected capabilities or their recorded source paths. The source edge chain preserves implemented, proposed and editorial scopes; the group link adds navigation, not runtime integration.
- Product work & delivery ↗Service context for this group
Derived through explicit member products, selected capabilities or their recorded source paths. The source edge chain preserves implemented, proposed and editorial scopes; the group link adds navigation, not runtime integration.
- Product feedback & learning ↗Service context for this group
Derived through explicit member products, selected capabilities or their recorded source paths. The source edge chain preserves implemented, proposed and editorial scopes; the group link adds navigation, not runtime integration.
- Billing, finance & measurement ↗Service context for this group
Derived through explicit member products, selected capabilities or their recorded source paths. The source edge chain preserves implemented, proposed and editorial scopes; the group link adds navigation, not runtime integration.
- Usage, metering & metrics ↗Service context for this group
Derived through explicit member products, selected capabilities or their recorded source paths. The source edge chain preserves implemented, proposed and editorial scopes; the group link adds navigation, not runtime integration.
The source behind this role.
Repository evidence identifies the implementation and its scope. Deployment and host qualification are separate.
- packages/platform/agents/harness/README.md ↗
Controlled inspection and verification with assertions, proof artifacts, replay and canonical quality receipts.
Evidence quality is explicit; a retry or fixture run does not qualify a production environment. - packages/platform/agents/agent-evals/README.md ↗
Candidate versus baseline trials, graders, replay, safety gates and progressive qualification.
Simulated evaluation does not grant effects; live qualification requires exact health evidence at every checkpoint. - packages/platform/agents/agent-evals/src/model-qualification-lifecycle-v2.ts ↗
Promotion checks the simulated qualification, production canary, exact digest binding and nonempty evidence before producing live_verified material.
The helper verifies supplied material; this audit did not observe a production canary or actual qualified deployment.
Other services in this layer.
Opportunity research & evaluation
Develop and compare company ideas using captured research, explicit evidence, hard gates, economics and portfolio decisions.
Inspected reusable source · host qualification separateExplore the service role ↗Builder & product delivery
Turn product intent and source material into blueprints, implementation plans, controlled candidate changes and handoff evidence.
Inspected reusable source · host qualification separateExplore the service role ↗Agents & coordinated execution
Compose agent sessions, tools, worker fleets and execution lanes with bounded delegation, checkpoints and evidence.
Inspected reusable source · host qualification separateExplore the service role ↗CI, release & progressive delivery
Execute reviewed quality plans, collect build evidence, coordinate release checkpoints and support bounded rollout and rollback.
Inspected reusable source · host qualification separateExplore the service role ↗Runtime operations & recovery
Plan and execute qualified infrastructure changes, observe drift and retain exact receipts for reconciliation, recovery and rollback.
Inspected reusable source · host qualification separateExplore the service role ↗Governed improvement & self-healing
Turn operational signals into scoped candidates, independent evaluations, controlled canaries, observed promotion and recoverable rollback.
Inspected reusable source · host qualification separateExplore the service role ↗Policy, review & governance
Bind decisions, reviews, approvals, constraints and audit evidence to the exact proposed work and current scope.
Inspected reusable source · host qualification separateExplore the service role ↗Adaptive product surfaces
Compose product-owned marketing and application experiences from shared design packs, surface contracts, role-aware layouts and bounded personalization.
Inspected reusable source · host qualification separateExplore the service role ↗App submission readiness
ReviewReady audits supplied mobile-store evidence into reviewable readiness findings and rejection-response preparation.
Inspected reusable source · host qualification separateExplore the service role ↗