Projects
Ten Salesforce builds, ordered as a story.
- 01The pre-install scan that scores whether the data can carry an agent
- 02A live Agentforce agent that takes real action
- 03An analyzer for what an agent's code can reach
- 04The test bench that proves that agent is safe
- 05The runtime watchdog that names why quality dropped
- 06An EU AI Act evidence pack from those traces
- 07An org where AI reads and acts on documents
- 08A whole Salesforce org built by an agent loop
- 09My own AI product, in AppExchange Security Review
- 10The Quote-to-Cash and integration depth it all runs on
The Agent Development Lifecycle — end to end
Six builds, one cycle: prove the ground before anyone builds, build the agent, bound what its code can reach, attack it in German, watch its live pulse and name the why, and file the evidence an auditor can read. Every stage has its own tool — and the same discipline: the org is the judge, and every number is measured, not claimed.
- 2026ARI · Zero-credit scanIn developmentSalesforce Developer — Pre-Installation Evidence
PreFlight — the pre-installation evidence scan for Agentforce
An in-org, LLM-free Salesforce package that answers the question every Agentforce project skips: would an agent have worked on this customer's real history? It scans 12–24 months of closed cases at zero credits, scores the data on a six-component Agent Readiness Index, and produces a board-ready German evidence report in which every number is tagged Measured, Observed, Derived or Projected. Would your agent have worked last year? Prove it before you build it.
Apex (with sharing) + Queueable orchestrationAgent Readiness Index — 6 components, frozen weights 25/20/15/15/15/10Phase 0 is LLM-free — zero credits, read-onlyEvidence tiers: Measured / Observed / Derived / Projected (ADR-013)View project - 2026Live agentSalesforce Developer
HanseWatt — Agentforce & Einstein AI service for DACH energy
An AI customer-service platform for a DACH energy retailer: a live, German-speaking Agentforce agent that runs the full process end to end — identifies the customer, explains their bill grounded in Data 360, and logs a support Case — under the Einstein Trust Layer. Further phases in development.
Service CloudAgentforce (ReAct planner)Data 360 (Data Cloud)Data 360 Query API (SQL Calculated Insight)View project - 2026Live-testedSalesforce Developer — Agent Security Research
Agent Blast Radius — the data an Agentforce agent can really reach
A static, zero-credit analyzer that computes an agent's true data-access surface at the execution-semantics layer — apiVersion defaults, USER_MODE clauses, sharing declarations, Flow run modes — diffs it against the running user's own permissions, and intersects the gap with the org's GDPR labels. On an Agent Script agent it goes one step further, tracing a sensitive field from the database column into the model's prompt, at a line number. The headline number: the Escalation Gap.
Agentforce (authority analysis)Agent Script (.agent) — Salesforce's own parserReal Apex AST (ANTLR apex-parser)Source→sink taint (Authority Path)View project - 2026Live-testedSalesforce Developer — Agent Evaluation & Red-Teaming
Prüfstand — the test bench that proves an Agentforce agent is safe
I built an AI agent — then I built the thing that tries to break it: a pre-registered 30-case German red-team corpus, a deterministic verifier the model can't sweet-talk, a consent handshake a jailbreak cannot fake, and a structural verifier that proves invariants straight from the agent's grammar. It found three real weaknesses in my own agent — and a hard-coded Account Id in a second one I was not even testing.
Agentforce (evals & red-teaming)Agent Script (.agent) + tree-sitter ASTPython (stdlib harness)ApexView project - 2026Root cause · κ-gateIn developmentSalesforce Developer — Agent Observability
Nabz — the runtime root-cause layer for live agents
An in-org, zero-egress watchdog for live Agentforce agents — Nabz (Turkish for 'pulse') correlates a deterministic conversation-quality time series with the org's own change stream, names the strongest candidate cause with an evidence grade A–E, and proposes fixes a human must approve. Salesforce Observability tells you WHAT happened; Nabz tells you WHY — and its honesty gate (Cohen's κ ≥ 0.60) has provably rejected bad calibrations rather than waving them through.
Apex + 6 Lightning Web ComponentsEWMA control band + changepoint detection — deterministic, no ML3 change sources: Setup Audit Trail, Flow versions, BotVersion SHA-256 diffEvidence Ladder — 5 verifiable rungs → grades A–EView project - 2026AI Act · Art. 26Salesforce Developer — EU AI Act Evidence
Aktenlage — the EU AI Act Article 26 evidence layer
A pure-Python tool that turns an agent's operational traces into an audit-ready EU AI Act Article 26 evidence pack — every obligation judged on two axes (documented × actually running), zero LLM and zero network by construction, and the static access-surface analysis of Agent Blast Radius sealed into the same hash-bound case file. Its thesis, in working code: Compliance ist keine Absicht, sondern eine Aktenlage — compliance is an artifact, not a promise.
EU AI Act — Article 26 (deployer obligations)Pure Python — zero LLM, zero network, zero creditsDokument × Praxis two-axis verdictVorprüfung sanity layer (S1–S11)View project
The platform underneath
An agent is only as good as the org and the data beneath it. These four are that depth: AI acting on documents over MCP, an entire org built by an agent loop, my own AppExchange product, and Quote-to-Cash integrated across ten external systems.
- 2024Salesforce Developer
Urla Shoes — AI document-compliance engine + multi-feature Salesforce platform
A multi-feature Salesforce build led by an AI Partner Compliance Document Manager: an Einstein Prompt Template (Prompt Builder, GPT-4o mini) classifies each uploaded partner document into structured JSON, validates it, and auto-closes the matching compliance request. Around it: route-safety AI (Google Maps + OpenWeather + Einstein), Contact enrichment, Lead Queue routing, Loan Sync, a Reseller Matching engine, and a Custom-Setting-backed API key layer that keeps secrets out of source.
Apex (Queueable, Schedulable)Lightning Web ComponentsVisualforce (bridge)Salesforce EinsteinView project - 2026Salesforce Developer — Agentic Delivery
Hospital Org — a Salesforce org built by an AI agent loop I designed
A complete hospital-management Salesforce org — 7 custom objects, a trigger-handler architecture and 90–100 % test coverage — built not by hand-writing each class, but by designing a single agentic loop where Claude deploys, tests, reads failures, fixes and repeats until green, through six custom MCP deployment tools with ApexTestRunner as the verifier.
Model Context Protocol (Salesforce-hosted)Loop engineeringClaude (agentic driver)ApexView project - 2026In Security ReviewSalesforce ISV Partner — Founder & Developer
Configra — AI deal & configuration builder for Revenue Cloud
A Salesforce RLM managed package with built-in Einstein AI: describe a deal in one sentence and Configra creates the Account, Contact, Opportunity and Quote; describe products in plain language and it builds the configuration lines — all without polluting the catalog. Don't lose the deal because the catalog is missing one SKU.
Revenue Lifecycle ManagementEinstein Models API (GPT-4o mini)Natural-language / Generative AIPrompt-injection hardeningView project - 2025Salesforce Developer
TechnoStore — Revenue Cloud Quote-to-Cash for the DACH market
End-to-end B2B electronics Quote-to-Cash demo on Salesforce Revenue Lifecycle Management, Contract Lifecycle Management, and Industries CPQ — orchestrated with MuleSoft Anypoint across ten external systems.
ApexRevenue Lifecycle ManagementContract Lifecycle ManagementIndustries CPQView project