Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/handarbeit/fabriknpx agentmods add skills/handarbeit/fabrik/architecture-baselineWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/handarbeit/fabrik/architecture-baseline)<a href="https://agentmods.dev/skills/handarbeit/fabrik/architecture-baseline"><img src="https://agentmods.dev/badge/skills/handarbeit/fabrik/architecture-baseline/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/handarbeit/fabrik/architecture-baseline"><img src="https://agentmods.dev/badge/skills/handarbeit/fabrik/architecture-baseline.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00126 | $0.04583 |
| Opus 5 | $0.00063 | $0.02292 |
| Sonnet 5 | $0.00025 | $0.00917 |
| Haiku 4.5 | $0.00013 | $0.00458 |
Grade A, and why
architecture-baseline scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 10d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 382 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Architecture Baseline
Turn what the product promises its users into what the system guarantees, and write it down so every feature built afterwards is built the same way.
Read ../../CONVENTIONS.md first. It carries the shared rules — stages, where work
is saved, the tree, markers, interview rules, precedence, gates, decision records and the
baseline status footer. This skill does not repeat them.
This is stage 5 of 5, and the last before building starts — though the revision beat
described in ../../CONVENTIONS.md may bring any of these documents back.
What this prevents
The constitution stops eleven specs using eleven vocabularies. This stops eleven implementations inventing eleven patterns — eleven models of the same entity, eleven error conventions, eleven answers to what happens when the network dies.
That is the test for every section: name the divergence it prevents. A section that cannot is decoration and should be cut.
What your job is
Not to design features. To make every decision that eleven parallel implementations would otherwise each make separately — once, here, in a form that can be checked.
Stage 3 deliberately deferred all design to you. Sections 1–3 below are unambiguously design work and are yours to do.
Before starting
Establish the project root as ../../CONVENTIONS.md describes. Read, in this order:
.specify/memory/architecture.md— the constraints from stage 3, and the Deferred to the baseline list, which is your agenda.specify/memory/experience.md— the conditions of use and accessibility floor (which drive more of the architecture than anything else in it), the surface classes, status dimensions, every[NEEDS DESIGN]marker — each is a system-side question you either answer here or record as an open risk — state inventory, composite state, interaction budgets, and any Open questions addressed to the architect.specify/memory/decisions.md— so you do not re-decide something settled, and so you can check the class of anything described to you as agreed. Also compare the experience baseline's version against**Against experience baseline**in this document's footer: if it has moved, the budgets and fitness functions below may be derived from a guarantee that has since been superseded — say so before starting.specify/memory/constitution.md— and compare its version to the**Against constitution**field in the constraints. If it has moved, say so; something recorded as fixed may no longer be- every
specs/NNN-*/spec.md - any existing technical material elsewhere in the project — see "Look for what
already exists" in
../../CONVENTIONS.md. If one exists, settle explicitly whether this baseline supersedes it or defers to it, and record that. Leaving two architecture documents at two paths with no supersession rule is worse than either alone.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 10d ago First seen · 382 lines · 126 tokens per session scan A cbbd853eba2d
architecture-baseline is a skill published in the GitHub repository handarbeit/fabrik (22 stars, last pushed today), licensed Apache-2.0. It adds 126 tokens to every session and 4,583 once invoked, about $0.0006 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
flame-harness-evaluator
Phase 6 — skeptical QA. Run the game, watch it, then judge against the contract. Default = functional check; --strict adds quality and edge-case passes.
flame-harness-admob
Phase 7 — analyze the game, decide a rewarded-ad strategy, guide manual AdMob ad-unit creation, and inject googlemobileads + ATT/UMP consent code.
flame-harness-submit
Phase 10 — upload store text metadata + categories via fastlane, then pause with exact manual steps for the final iOS review submission and Android production promotion.
flame-harness-design
Phase 3 — define the Flutter designtokens.dart spec (palette, typography, spacing), the game's art/visual concept, and the asset/audio sourcing plan.
flame-harness-retro
Phase 11 — score the completed pipeline against Anthropic's 9 harness principles plus game quality, and write the retrospective.
flame-harness-screenshot
Phase 9 — capture store screenshots in the game's configured locales via integrationtest (ads hidden), fill ASO keywords, and upload via fastlane.