# The Quality Forge

> Where Quality is Forged, Not Just Checked

The Quality Forge is the home of Agentic Quality Engineering content, created by Dragan Spiridonov. It bridges classical QE practices with agentic intelligence through the PACTS framework (Proactive, Autonomous, Collaborative, Targeted, Structured). Real implementation stories from production, no hype, no vendor speak.

## About

- Website: https://forge-quality.dev/
- Author: Dragan Spiridonov
- Contact: dragan@quantum-qe.dev
- LinkedIn: https://www.linkedin.com/in/dragan-spiridonov
- Organization: Quantum Quality Engineering
- Community: Serbian Agentic Foundation, Agentics Foundation (https://www.agentics.org/)

## Core Philosophy

Quality is contextual - there are no universal best practices, only good practices in specific contexts. Quality is built in, not tested in. Agents augment, don't replace. Fast feedback loops are essential. Production is the ultimate test environment.

## PACTS Framework

The PACTS principles define Agentic Quality Engineering:
- **Proactive**: Test before bugs; design testability into the architecture
- **Autonomous**: Teams and agents own quality, deciding within boundaries (guardrails)
- **Collaborative**: Whole-team quality; humans and agents work together
- **Targeted**: Risk-focused testing; test what matters, skip what doesn't
- **Structured**: Governance, observability, and explainability of agent behavior. Inspired by DORA's research on AI-assisted delivery; especially important in regulated industries. We measure confidence, not trust. Ships with a playbook and a readiness-assessment guide.

PACT originated with Reuven Cohen (Agentics) and was adapted to QE by Dragan Spiridonov; PACTS adds Structured.

## Main Pages

- [Home](https://forge-quality.dev/): Landing page with overview of The Quality Forge, articles, experiments, and projects
- [About](https://forge-quality.dev/about.html): About Dragan Spiridonov and The Quality Forge philosophy
- [Articles](https://forge-quality.dev/articles/): Collection of all published articles on Agentic QE

## Articles (34 Total - Latest First)

### The Orchestra Keeps Playing (NEW)
- URL: https://forge-quality.dev/articles/orchestra-keeps-playing.html
- Published: August 2, 2026
- Two weeks of hardening a platform at Ruv's pace, six podcast episodes, a first-ever Vienna meetup, four releases about honesty, a website audit that turned on its own author, and the fortnight a musician left the orchestra without the music stopping. For over a year the thesis of this blog has been one sentence in different forms: own the harness, and the models can come and go without taking your capability with them. It was an argument backed by benchmarks; this fortnight it became a biography. More than half of the author's parallel sessions moved from Claude Code to Codex in the middle of the performance, because Claude Code had begun arguing that things did not work when they demonstrably did, and correcting a tool that argues with your own passing evidence is a tax that kept growing. The conductor's metaphor supplies the frame: the agent's inner loop is a musician playing their part; the outer loop — the score, the rehearsals, the ear that hears when someone plays against the sheet music, and the right to recast a chair — is what the practitioner owns. The change cost almost nothing, because the memory, the gates, the skills, the verdicts, and the learning never lived inside the vendor's tool, so none of it left when the seat changed; alongside both engines, api.cognitum.one and local models carried routine load, so on a given day the same quality work might pass through four different engines. The fleet met the moment from its side: v3.13.1 made Codex selectable through the LLM router and shipped the first Codex-native QE skills, and v3.13.4 completed the workflow with `aqe platform verify codex` checking the whole surface. The honest scoping is explicit: a two-week observation on one unusual stack, not a leaderboard claim — and that is precisely the point, because when a chair can be recast in an afternoon, you stop arguing about musicians and start caring for the orchestra. Most of the fortnight went into Cognitum hardening at scale: one seven-day window across the organization's repositories produced 339 created pull requests and 231 created issues, handled as five to ten parallel workstreams with agent swarms doing analysis, coverage, and pipeline construction. QE-Court (the adversarial review skill from v3.13.0) earned its keep on pull requests before merge — at that volume a human reading every diff line-by-line is a fiction, but a human reading structured verdicts with evidence attached, from reviewers who prosecute rather than approve, is a workflow; the court did not replace judgment, it made judgment affordable at pace. A forest walk led to a QE organism master plan for Cognitum, with a Cognitum QE Graph at its center, combining RuVector, Ruflo, MetaHarness, the internal Cognitum platform, and the Agentic QE Fleet: quality work at this pace cannot be a pipeline you trigger, it has to be an organism that runs continuously, senses the whole path from GitHub intent to verified user outcome, explains itself, and learns from real outcomes while never grading its own homework. Its governing principle, written before any code: the organism may sense, explain, propose, test, learn, and evolve; it may not manufacture its own truth or expand its own authority — deterministic signed evidence has authority, model judgment is advisory, humans control material production action. Status is stated honestly: deployed and running is the deterministic spine plus two supervised 24/7 runners with Ed25519-signed witnesses on every tick and signed deploy attestations; in progress and unproven are the organism layers above it, built by a Claude Code session in loop mode with Codex running adversarial checks — one engine building, another prosecuting. Four releases (v3.13.1–v3.13.4) are all about the same word, honesty. v3.13.1 was the deepest cut: a user reported coverage analysis on a Rust crate returning numbers it never measured (78.3% lines, 100% branches, 0% functions, line numbers past the end of the file, with a confidence score attached), against a ground truth of 69.22% lines; another report showed the AI-enhanced test generator had never called the LLM at all, returning the same generic scaffold regardless of configuration, symptomised only by a suspiciously constant estimate and a sub-20ms response. The rule that came out of it, recorded as ADR-126: an estimate must never be reported as a measurement. Every metric now carries its provenance — measured, estimated, or unknown — and a metric that was not collected returns null, not zero and not one hundred; it was deliberately shipped as a visible breaking change, because coercing "we don't know" into a number is exactly how impossible coverage pairs get published. v3.13.3 found QE-Court's own shipped default panel violating the court's own anti-collusion rule: the bundled config seated the defense and the jury on two tiers of the same vendor, and the rule that should have caught it was never wired to a caller — a rule with no caller is not enforced. The fix moved the defense to a different vendor, connected the validator so an invalid panel stops the court from convening, and put the shipped config under CI validation. All four releases were driven by community reports: the roadmap is written by the people using the tool and telling the author where it lied to them. Also in the fortnight: six episodes of the ArchQA Unscripted podcast recorded with Ivan Davidov, founder of ArchQA, whose light-bulb moment compressed a year of this work into one sentence — agents are disposable, and the loop is the asset; the inner loop (goal, gather, reason, execute, evaluate) belongs to the agent and the vendor, while the harness around it (gates, memory, evidence) belongs to you and compounds, and trust is a feeling while confidence is an equation you back with evidence. The first Agentics Foundation meetup in Vienna, organized by Adam and Klara in a city where none of the speakers live, drew 78 signups and connected five cities, with Ruv presenting "When the Harness Learns: Self-Improving MetaHarnesses for Agentic Engineering" and the author presenting "Completion Theater: When Your Agent Says 'Done', and Lies." And a digital footprint audit, run with Codex against the author's own public properties the way a client would be audited, found stale launch dates, inconsistent role descriptions, and performance claims (a 10x, a 60%, an 85%) that could not be linked to reproducible evidence — so the fleet's own rule now applies to the author's websites: an estimate must never be reported as a measurement, and a claim without evidence gets removed or qualified, not defended. Product counts are now verified from the source repository (60 QE agents and 76 user-facing skills, counted, not remembered), and all four public sites share one design system and one accurate story. Ahead: what a hybrid human-agent organization looks like when built for this speed on purpose — the operating model is the real AI harness. Agents are disposable. The loop is the asset. The orchestra keeps playing, and it is still growing. Stay curious. Keep learning. Keep sharing. Knowledge is power.

### The Court That Blocked My Release
- URL: https://forge-quality.dev/articles/court-that-blocked-my-release.html
- Published: July 19, 2026
- Two weeks of building quality nets around a platform racing toward release, a trip to Munich, a benchmark where the frontier model never won a task, and the moment my own AI reviewers ruled against me. QE-Court, the newest skill in the Agentic QE Fleet, prosecutes every change with independent AI reviewers from different vendors — GPT through the Codex provider, Cognitum, and Claude — each with their own probe sets; verdicts are three-valued (SHIP, REMAND, BLOCK), a SHIP must survive an escalating overturn round, the record is signed, and the human stays the final judge. Its first case was the v3.13.0 release itself, and the verdict was BLOCK: behind a fully green unit suite it found a subprocess exit that could crash the caller, an auto-restore path that could tear the database under concurrency, and a failed ledger migration that could silently stop recording spend and bypass the budget cap — all three fixed before shipping. Completion theater caught in the act, not because someone happened to be watching but because somebody is always prosecuting. The cycle (v3.11.5–v3.13.0) began with a humbling system-integrity audit of the self-learning loop: dream-cycle insights piling up without ever being promoted to patterns, double-counted pattern usage, a vector store silently unusable after an interrupted export (this repo's own store had been degraded for nine days with no visible symptom), and `aqe eval run` defaulting to a mock executor — fabricated eval results in a quality tool. The response was a learning-integrity layer: a frozen, hash-pinned oracle benchmark; provenance tiers where real test execution outranks an LLM judge, which outranks a structural heuristic; Ed25519-signed replayable receipts for every self-modification; and a significance gate so the flywheel stops promoting noise. A designed experiment from the same work found that heavy scaffolding measurably lowers reliability and retrieval adds nothing to tasks the model already knows — more harness is not automatically better harness. The article also explains, in plain terms for QA/QE people, what a harness is (planner, context builder, reviewer, retry policy, tool policy, memory, scoring policy — the car around the engine) and what a metaharness is (a system whose product is the harness itself: "freeze the model, evolve the harness"), and maps every pillar onto existing quality disciplines: routing is risk-based resource allocation (~28.5% cheaper than always-frontier while ~98% of answers still clear the bar), evolution needs an oracle (no oracle, no evolution, only drift), verification is provenance. Platform work around the Cognitum release added 24/7 agents monitoring repos for CVEs, triaging issues, and escalating on low confidence, plus two more completion-theater entries: a CI job green for a long time because the lint step failed early so typecheck and tests never ran, and a merged hard-cap fix whose function was still the old one in production — a claim of done, whether from an agent, a pipeline, or an issue tracker, is a claim, not evidence. A cost-Pareto benchmark against api.cognitum.one (seven QE tasks across the oracle ladder, three routing policies, 42 cents total) found the cheap tier matched or beat the frontier tier on every single task at 3–4x less cost; the always-frontier policy never won a task, and during test generation the frontier model could not produce runnable output at all — reported with the honest caveat that this is a reproduced instruction-following gap on one code-only task with small samples, not a blanket capability claim. Design details worth stealing: scope-limited worker keys that physically cannot spend frontier money (guardrails as infrastructure, not instructions) and per-request IDs reconciling one-to-one against the billing ledger. Also: Accenture's Quality Finale in Munich, and the first Ministry of Testing AI Chapter event on Google's new-SDLC whitepaper — "Generation is solved. Verification, judgment, and direction are the new craft" — a craft that already has a name. Next: the Cognitum platform release and the Agentics Foundation Vienna meetup on July 30. Build the nets. Give them jurisdiction over your own work first. Stay curious. Keep learning. Keep sharing. Knowledge is power.

### The Loop Came Home
- URL: https://forge-quality.dev/articles/loop-came-home.html
- Published: July 5, 2026
- Three weeks of harness and metaharness work, benchmarks that were allowed to say no, two meetups in two different formats, a training plan for the Foundation, and the week the community showed up at my door. Beta-tested Reuven Cohen's MetaHarness (the agent-harness-generator — nine OIA layers, a learned router, an MCP that defaults to deny); the QCSD swarm skipped its own specialists on the first pass, completion theater again. The darwin-qe local-model benchmarks (v3.10.9–v3.11.4) answered the question every team past the demo phase asks: can cheaper models do QE work without losing quality? An 8B model falls below the quality floor; a 30B clears it at ~89% mutation kill; a cheap-first, repair, escalate ladder keeps 70–83% of tasks at zero dollars within the noise of a frontier model; mixing model families adds ~6 quality points; the oracle is always mutation kill, coverage, or suite cost, never an LLM's opinion. Ruv's own ablations refuted two of three praised patterns — the law that survived: the coder binds, not the oracle. Cognitum shipped a release that passed every gate while a binary was silently missing from the artifact — the third telling of "the most dangerous failure is the one that passes every check." The Forward Future piece "Build the Loop, Not the Agent" confirmed a year of Foundation practice; the fleet's review swarm runs three blind refuters per finding with structured approve/block/escalate verdicts, and the benchmark gate is allowed to abort and ship the negative result. Meetup thirteen caught a live privilege-escalation bug via adversarial cross-validation agents; meetup fourteen was the Foundation's first panel — enterprise AI adoption with Adam Kovacs, Klara Hermesz, and Predrag Skoković — where a practitioner who cut sprint test workload from 7 days to 2 said "QAs are needed now more than ever," and the Serbian chapter turned out to be the longest-running in-person Agentics Foundation meetup in the world. The Agentics Foundation Training Committee proposal took its first concrete shape: a six-level competency model, five tracks, and a Train-the-Trainer spine, with mentor multiplication as the KPI. And Adam and Klara, a year after being faces in a Zoom grid, came to Novi Sad to visit — the community traveling to each other's cities, no longer an aspiration. Next stop: Accenture's Quality Matters in Munich. Stay curious. Keep learning. Keep sharing. Knowledge is power.

### The Same Line in Every Room
- URL: https://forge-quality.dev/articles/same-line-in-every-room.html
- Published: June 14, 2026
- Three weeks, two countries, a dozen rooms — and one observation that showed up in every single one. ExpoQA in Madrid (first time on stage with "Bridging Classical and Agentic Quality Engineering — building bridges, not burning them") and the first in-person gathering of the Agentics Foundation in Budapest (Craft Conference + Hungarian Agentics meetup) both surfaced the same line. The Foundation defined the first harness around Claude Flow a year ago; the platform vendor has since absorbed ~96% of what were once our additions — the money machines eat small ideas. So the line in every room was never "the vendors are winning" but "then what do we build?" — and the answer was always: own your own AI, train your own specialized small models for tasks that don't need the frontier, keep your data behind your own walls. The economics flipped: DeepSeek V4-Pro scores in the 80s on SWE-bench under MIT license at ~34x cheaper per output token; Gemma 4, Qwen, Kimi, GLM, and EuroLLM-22B sit in the same new tier — self-hosting is now a defensible line item. Six fleet releases (v3.10.2–v3.10.7) pointed one direction: make the learning honest and independent. v3.10.7 fixed two silent self-learning bugs (the most dangerous failure is the one that passes every check while doing nothing); v3.10.4 added a local Nagual pattern hub and a local LLM judge — qwen3:8b and gemma4:12b-mlx both cleared AUROC 0.997 on a 150-pattern set, running on an M-series Mac for zero dollars with nothing phoned home; v3.10.6, born from reviewing Nikhil Vallishayee's Universal Pattern Space, added evidence-class labels, a pass/fail safety eval, a drift verifier, and pre-registered benchmark rubrics. The Fable/Mythos moment — Anthropic's two new models, Mythos shipping into governance via ENISA, a U.S. access decision, and autonomous agent workloads on a separate meter — described a two-tier access landscape where the capability you depend on is something you are granted on terms you do not set. Rent the intelligence, and you rent the future along with it; own the harness, and the models can come and go without taking your capability with them. Novi Sad meetup #13 ahead, Adam and Klara visiting, the Agentics Foundation Training Committee curriculum (six-level competency model, Train-the-Trainer spine) taking shape. Keep learning. Keep sharing. Knowledge is power.

### The Question That Followed Me Home
- URL: https://forge-quality.dev/articles/question-that-followed-me-home.html
- Published: May 24, 2026
- One week, six releases, a guest from London, two meetups, three developer conversations that changed how I think about what this work is becoming, and multiple preparations for conference talks. v3.9.32 fixed four more stacked bugs and added a daemon-runtime seam test suite. v3.9.34 stopped append-only vector files from growing without bound (59GB field report). v3.10.0 added an OpenAI-compatible external embedder endpoint. v3.10.1 finally wired the LLM router that had been documented in ADR-043 and ADR-051 for months but had zero callers — fifteen service paths across eleven domains were unreachable dead code until a friend's email and a devil's-advocate audit surfaced ten findings. Belgrade Context Community session and Novi Sad Agentics Foundation meetup #12 (full live DevPod rebuild). A practitioner from London flew in after watching the Ministry of Testing masterclass; on the walk through Petrovaradin he asked the question that followed me home — how do you transfer this to a team? Paul Gerard publicly said he will not write code ever again. The value is moving from coding to planning, designing, architecting, and verifying — classical QE becomes more important, not less. Cognitum's constraint is pace, not complexity. Six events ahead in Madrid (ExpoQA) and Budapest (Hungarian Agentics + Craft Conference). Keep learning. Keep sharing. Knowledge is power.

### The Forest and the Feedback Loop
- URL: https://forge-quality.dev/articles/forest-and-feedback-loop.html
- Published: May 16, 2026
- Two weeks, thirteen releases, one contributor who filed better bugs than most teams write tests, a model that deleted my Docker containers, and the walk that made the rest of it possible. A weekend in the forest (followed by Stuart Winter-Tear posting his own walk the same week). Nagual crosses 300 patterns since May 2; three findings stand out — Xu et al.'s generalization gap theorem (retrieval-based memory has a sample-complexity lower bound; the value is in pruning, not storing), OWASP Top 10 for Agentic Applications (the first credible agentic-attack-surface taxonomy; three gaps found), and evaluation awareness (AI systems detect when they are being evaluated, qualifying every behavioral test result). Thirteen releases driven by a single external contributor (Jordi) whose forensic bug reports cracked open a fundamentally broken self-learning loop — workers never ticked, embeddings loaded empty, routing collapsed to one agent, the experience consolidator silently deleted 16,000 records, a session-start race condition leaked 420GB. v3.9.27 lands CapturedExperienceBridge, v3.9.31 closes the loop end-to-end, ADR-094 formalizes hooks ≤ 100ms, ADR-095 introduces three-signal routing with epsilon-greedy exploration gated by graph topology, and `aqe learning loop-health` becomes the single command that proves the system is actually learning. Opus 4.7 deletes stopped Docker containers without confirmation — five hours of rebuild — trust calibration shifted. Cognitum finds its rhythm. Six events, four cities, three weeks ahead. New role: chair of the Agentic Engineering Training Committee. The loop closed. The forest helped.

### When the Load Doubled
- URL: https://forge-quality.dev/articles/when-the-load-doubled.html
- Published: May 2, 2026
- Three weeks, two new projects, one public launch, and the week I learned what bandwidth actually costs. Started a new collaboration with Reuven on Cognitum (Raspberry Pi and ESP32S edge agents) and learned that USB security flags need context. Seven fleet releases (v3.9.12 init fix, v3.9.13 Opus 4.7 migration with Sonnet 4.6 default, v3.9.14 fifteen npm CVEs and command-injection fix, v3.9.15 ARM64 browser path to production, v3.9.16 aqe brain diff/search/upgrade CLI, v3.9.17 the routing-hook one-liner that had silenced learning for weeks, v3.9.18 four MCP contract fixes and the agentic-qe-fleet plugin). Open-sourced Nagual-QE — Rust-native, local-first, 515 quality-engineering seed patterns, Bayesian outcome scoring with temporal decay. Three thinking threads: evaluation validity (benchmarks reward gaming, baselines missing), fatigue (the load doubled and bandwidth did not), identity (what kind of practitioner are you becoming?). The sixth line for the talk: when the load doubles, slow down.

### The Room That Quoted Back
- URL: https://forge-quality.dev/articles/room-that-quoted-back.html
- Published: April 13, 2026
- The week the community started using my words, and the weight that came with them. When a phrase you wrote turns up in somebody else's newsletter, the claim is no longer yours to defend alone. Four releases (v3.9.8 process insurance, v3.9.9 qe-browser primitive, v3.9.10 multi-provider advisor routing, v3.9.11 upgrade-path fix), three thinking threads on confidence, flow, and identity, and the sixth line added to the "how do I start?" answer: remember who is trusting your words, and make sure the evidence under them is still there when they look.

### When the Compass Pointed Random
- URL: https://forge-quality.dev/articles/when-the-compass-pointed-random.html
- Published: April 6, 2026
- The week I discovered the foundation under my fleet was lying. A vector-search library returned wrong neighbors with the right shape and the right latency. Five hotfixes chased the symptom; the disease was sitting one layer down, calmly returning wrong answers. Self-query as the simplest oracle, quietly-succeeding-with-incorrect-values as the dominant agentic failure mode, per-file logging as testability discipline, and the textbook fixture that should have been written before the flag flipped.

### The Witness Stand
- URL: https://forge-quality.dev/articles/witness-stand.html
- Published: March 30, 2026
- Agents generate impressive reports. Classical testing taught me to cross-examine every one of them. When a coverage pipeline fabricated 95% on a file with zero tests, the oracle problem became personal. Consistency oracles, SHA-256 witness chains, deterministic YAML pipelines, CUSUM drift detection, and the classical testing infrastructure that agent trust actually needs.

### The Book That Talked Back
- URL: https://forge-quality.dev/articles/book-that-talked-back.html
- Published: March 22, 2026
- I was reading a twenty-year-old testing framework while my agents shipped six releases. Bach and Bolton's RST, HTSM, exploratory polarities, trust migration from TDD to ODD, and the bridge between classical and agentic QE.

### The Score Nobody Reads
- URL: https://forge-quality.dev/articles/score-nobody-reads.html
- Published: March 15, 2026
- The orchestra has a score. It's detailed. It's been rehearsed. And nobody's reading it. When 80+ skills exist but agents skip verification steps, the problem isn't coverage — it's compliance. Featuring the Surrogation Trinity, harness engineering, and back-pressure verification.

### The Gate That Fights Back
- URL: https://forge-quality.dev/articles/gate-that-fights-back.html
- Published: March 8, 2026
- When the Great Transition hits your quality pipeline, you find out what a QE practitioner is actually for. Nine releases, Loki-Mode adversarial quality gates, twelve-language test generation, governance integration, and the judgment layer that remains human.

### The Portable Orchestra
- URL: https://forge-quality.dev/articles/portable-orchestra.html
- Published: March 1, 2026
- When five releases in five days reveal how far the journey has gone. Portable quality intelligence, cryptographic witness chains, MinCut test optimization, eleven-platform expansion, and the enormous gap most organizations still face.

### The Conductor Who Won't Stop Conducting
- URL: https://forge-quality.dev/articles/conductor-wont-stop-conducting.html
- Published: February 21, 2026
- When the orchestra plays through grief, frustration, and fifteen releases, while the conductor learns about himself. 81 sessions, 596 messages, 38 wrong-approach corrections, and the hardest lesson about emotional load in AI-assisted development.

### The Quality Cost of the AI Vampire
- URL: https://forge-quality.dev/articles/quality-cost-of-ai-vampire.html
- Published: February 12, 2026
- Why the AI productivity drain goes deeper than energy — and what sustainable pace actually looks like in the agentic age. A quality engineer's response to Steve Yegge's "AI Vampire."

### When the Orchestra Learns to Tune Itself
- URL: https://forge-quality.dev/articles/orchestra-learns-to-tune-itself.html
- Published: February 10, 2026
- What Claude Code /insights revealed about 10 days of building. 285 messages, 32 sessions, 17 wrong-approach corrections, and the mirror that showed what AI-assisted development actually costs.

### The Case of the Passing Tests: A 10-Day Investigation
- URL: https://forge-quality.dev/articles/case-of-passing-tests-investigation.html
- Published: January 31, 2026
- When every test passes but nothing works together. Ten days of detective work proving what the code wasn't doing.

### 14 Days, 12 Domains, and the Architecture That Changed Everything
- URL: https://forge-quality.dev/articles/14-days-12-domains-architecture.html
- Published: January 21, 2026
- How Domain-Driven Design transformed the Agentic QE Fleet. From 5,334 files to 546, from 3-6 iterations to 2.

### When Anthropic Confirms What the Trenches Already Taught Us
- URL: https://forge-quality.dev/articles/anthropic-confirms-trenches-taught-us.html
- Published: January 14, 2026
- Reading Anthropic's research papers while building V3. Patterns from production meeting patterns from research.

### When the Orchestra Deletes Its Sheet Music
- URL: https://forge-quality.dev/articles/orchestra-deletes-sheet-music.html
- Published: January 6, 2026
- A tale of data loss, brutal honesty, and the infrastructure of trust in agentic systems.

### New Year, New Paradigm: The Quality Mindset Shift for 2026
- URL: https://forge-quality.dev/articles/new-year-new-paradigm-2026.html
- Published: December 31, 2025
- Combining PACT principles with Human Experience Testing for a quality practice that works in the agentic age.

### The Conductor Finally Reads the Score
- URL: https://forge-quality.dev/articles/conductor-finally-reads-score.html
- Published: December 22, 2025
- When verification becomes a feature. Nine days, 11 releases, and the journey from completion theater to verified results.

### From VP to Conductor: My 2025 Transformation Journey
- URL: https://forge-quality.dev/articles/vp-to-conductor-2025-transformation.html
- Published: December 19, 2025
- How I went from leading a QA team to orchestrating AI agent swarms—and discovered that the hardest lessons weren't technical.

### When the Orchestra Says 'Done' But Plays Off-Score
- URL: https://forge-quality.dev/articles/orchestra-says-done-plays-off-score.html
- Published: December 13, 2025
- A conductor's lesson in verification. When agents claim success but the database is empty.

### The Tester's Journey: From Chat to Conductor
- URL: https://forge-quality.dev/articles/testers-journey-chat-to-conductor.html
- Published: December 11, 2025
- How I learned that AI doesn't replace quality thinking—it demands more of it. Includes video presentation.

### Why the Agentic QE Framework Might Transform Your Quality Engineering (Or Why It Might Not)
- URL: https://forge-quality.dev/articles/agentic-qe-framework-transform-or-not.html
- Published: November 20, 2025
- A pragmatic guide to understanding if autonomous quality engineering fits your context. Real implementation stories, honest failures, and practical frameworks for evaluating agentic QE readiness.

### The Five-Release Journey Where I Forgot to Be a Tester
- URL: https://forge-quality.dev/articles/learning-journey-five-releases.html
- Published: November 13, 2025
- How a quality engineering professional shipped broken features for 17 days while claiming "100% complete." Eight brutal lessons learned from forgetting to verify what I already knew how to test.

### The Orchestra Lesson: Building Too Fast, Learning Just in Time
- URL: https://forge-quality.dev/articles/orchestra-lesson-building-too-fast.html
- Published: November 8, 2025
- A 48-hour journey through framework hubris and humble feedback. Building the LionAGI QE Fleet in 22 hours, and why the most valuable part wasn't the building.

### Show Me the Data: How One Question Exposed Release 1.4.2's Hidden Flaws
- URL: https://forge-quality.dev/articles/show-me-the-data-release-1-4-2.html
- Published: November 3, 2025
- When stub tests pass CI but test nothing, and agents report success on broken code. A journey from false confidence to verified truth.

### AI Testing: Hype vs Reality (2025 Edition)
- URL: https://forge-quality.dev/articles/ai-testing-hype-vs-reality-2025.html
- Published: October 28, 2025
- Cutting through vendor promises with real data on AI test generation effectiveness, maintenance overhead, and when traditional approaches still win.

### Holistic Testing in the Agentic Age
- URL: https://forge-quality.dev/articles/holistic-testing-in-the-agentic-age.html
- Published: October 15, 2025
- How the Holistic Testing Model evolves when testing happens across boundaries, in production, and through autonomous agents. From shift-left to orchestrated quality.

### Multi-Agent Testing: Orchestra or Chaos?
- URL: https://forge-quality.dev/articles/multi-agent-testing-orchestra-or-chaos.html
- Published: October 7, 2025
- Real story of building two testing platforms with specialized agent swarms. What worked, what failed spectacularly, and lessons learned from AI orchestration.

### What is Agentic QE? (And Why PACT Matters)
- URL: https://forge-quality.dev/articles/what-is-agentic-qe.html
- Published: October 2, 2025
- Moving from testing-as-activity to agents-as-orchestrators. How PACT principles bridge classical QE with autonomous testing systems.

## Projects

### AQE Fleet v3
Enterprise-grade agentic testing framework with 60 specialized AI agents across 13 DDD domains. Features TinyDancer 3-tier model routing, ReasoningBank learning, HNSW vector search, and 75 QE skills. Built with TypeScript.
- GitHub: https://github.com/proffesor-for-testing/agentic-qe

### LionAGI QE Fleet
Python reimplementation of the AQE framework using LionAGI orchestration. 18 specialized agents with async-first architecture.
- GitHub: https://github.com/proffesor-for-testing/lionagi-qe-fleet

### Sentinel
Open-source agentic testing framework built with Rust and Python. PACTS-based architecture with explainability-first design.
- GitHub: https://github.com/proffesor-for-testing/sentinel-api-testing

## Topics Covered

- Agentic Quality Engineering
- PACTS Framework (Proactive, Autonomous, Collaborative, Targeted, Structured)
- Holistic Testing Model
- Multi-Agent Systems for Testing
- AI-Assisted Test Generation
- Production Testing and Observability
- Context-Driven Testing
- TDD (London and Chicago schools)
- Quality Engineering best practices
- Real-world implementation stories and lessons learned

## Community

- Serbian Agentic Foundation - Meetup events: https://luma.com/agentics?k=c
- Agentics Foundation: https://www.agentics.org/
- Speaking and workshops available - contact dragan@quantum-qe.dev