All 27 built products. Every one is complete and validated; none is purchasable yet — checkout is pending (Phase 3) and this storefront is browse-only during pre-launch.
In the owner-approved launch order: agent infrastructure first, then the core harness/safety/quality systems, then the business operating systems.
A complete operating system for the context window: executable budget policies, compaction contracts with must-survive lists, retrieval caps, anchored context audits, and session-handoff manifests that keep long-running agents coherent.
A drop-in matrix that maps every tool an agent can call to a risk tier, an approval gate, and an audit requirement - enforceable by an orchestrator from one JSON file, deny-by-default.
Gives any persistent agent a governed memory layer: executable policies for what to store, what never to store, how long to keep it, and how to audit and prune it, with human review gates at every destructive step.
Standard recovery ladders for the most common agent failures: executable retry, back-off, escalate, and abort decision trees, with a failure-class taxonomy and incident records.
Compiles loose process descriptions into typed, gated, machine-executable workflow specs: every step with declared inputs, outputs, quality gates, approval gates, and a rollback path - plus task graphs and handoff manifests for delegation.
Turn loose prompts into reliable, reusable prompt harnesses with defined inputs, output contracts, quality gates, retry logic, and scoring rubrics.
Add human-approval checkpoints, permission boundaries, and rollback paths to any AI agent workflow before something breaks.
Score, compare, and improve AI output with repeatable scorecards, rubrics, and review workflows instead of gut feel.
A structured system to plan, build, quality-check, and ship digital products with AI assistance — from idea validation to launch checklist.
Templates, harnesses, and checklists designed to help write clearer, more complete, better-structured marketplace listings.
An agent-ready content production system for creators: plan, produce, repurpose, and quality-check content on a repeatable cadence.
A repeatable, AI-assisted marketing operating system built for local service businesses — weekly cadence, templates, and QA included.
Waves 2, 3, and 4: built, G10-approved (delegated review, 2026-07-07), prices owner-confirmed, launch positions 13-27 recorded. Not purchasable — like everything else here, until checkout exists.
Ready-to-run evaluation harnesses for agent behavior - task suites, anchored graders, and regression gates that tell you whether an agent change made things better or worse before it ships.
Layered, defense-only protection against prompt injection - untrusted-content quarantine policies, instruction/data separation rules, detection heuristics, and a defensive regression test suite.
Append-only audit trails for every class of agent action - schemas, generation rules, and review sampling plans that make agent behavior reviewable after the fact.
Build a structured safety case - claims, arguments, evidence - for any agent deployment, in a form both a reviewing human and an auditing agent can verify.
Runbooks for live agent incidents - classification, containment ladders, kill-switch procedures, and postmortem discipline, executable by the responding team and the agents themselves.
Package any agent capability as an installable, versioned, evaluable skill - manifest standards, trigger specs, resource packaging rules, test packs, and distribution-ready capability modules.
Rewrites MCP server and tool descriptions so client agents select the right tool the first time - with an anchored description rubric, a selection-failure catalog, and before/after test harnesses.
Turn a working assistant into a shippable product - configuration specs, knowledge packaging rules, guardrail checklists, and store-listing readiness in one kit.
Listings that buying agents can parse, score, and trust - a machine-readable listing standard with claim-evidence mapping and a completeness scorecard.
Machine-checkable entitlements for agent-consumed products - license records, seat and agent-instance rules, and access verification an agent can run before using what it bought.
Trust primitives for agent-to-agent commerce - seller verification checklists, claim-evidence standards, dispute records, and portable reputation schemas built on verifiable events, never simulated ratings.
Charter every agent in a multi-agent system - role, authority, boundaries, and escalation - as machine-readable contracts the agents themselves can check before acting.
Standard handoff envelopes for agent-to-agent and agent-to-human transfers - context, authority, open risks, and acceptance checks travel with the work.
A gate system that verifies an agent is production-ready - capability, safety, rollback, monitoring, and ownership checks with evidence required at every line.
Governs what a buying agent may evaluate and buy - budget rules, vendor criteria, approval chains, and purchase-record discipline as executable policy.
Prices marked "proposed" are revisable before launch; "owner-confirmed" prices are set but equally pre-launch. Bundles are on the bundles page; free samples on the samples page.