People
Practitioners and researchers whose published thinking is distilled into a practice BOSS ships.
-
Addy Osmanikey
AI code review; the 70%-to-80% problemgit-workflow feature-flags
-
Andrej Karpathy — the verifiability thesiskey
harness-engineering testing-with-agents
-
design-system
-
Dex Horthy — 12-factor agentskey
harness-engineering
-
Hamel Husain + Shreya Shankarkey
error analysis, evals-as-spec, judge validationtesting-with-agents
-
Jason Liu — systematically improving RAGkey
recall@k first; inventory vs capabilityretrieval
-
MonolithFirst (Fowler); ParallelChange (Danilo Sato, on Fowler's bliki)
ship-it-live scalable-architecture
-
schema evolution
data-schema
-
discovery vs documentation — why a written artifact is a means of deciding, not a deliverable
documentation
-
Michael Nygard — Documenting Architecture Decisionskey
the architecture decision record (2011) — the direct ancestor of BOSS's DEC-NNN: short, dated, superseded rather than editeddocumentation
-
Nathan Curtis — design tokenskey
design-system
-
the lethal trifecta; Agents Rule of Two; classifiers are non-deterministic; MCP prompt injection
agent-security mcp ship-it-live
Labs, companies & projects
Engineering writing, open-source projects and published guidelines BOSS builds on.
-
Anthropic — the Agent Skills spec and templatekey
skill-authoring
-
Anthropic containment + Redwood control
agent-security
-
effective harnesses for long-running agents; scaling managed agents; the agent-computer interface; contextual retrieval
harness-engineering retrieval documentation
-
retrieval
-
ConfigCat
flags as a vibe-coding safety netfeature-flags
-
Factory.ai
documented conventions rot; enforced conventions compoundscalable-architecture
-
GitHub Spec Kit / AWS Kiro
the spec-driven-development lineageharness-engineering
-
Google PAIR
ai-ux-patterns
-
GrowthBook
flag the model, not just the featurefeature-flags
-
METR
the developer perception gaptesting-with-agents git-workflow
-
Microsoft HAX guidelines
ai-ux-patterns
-
Nielsen Norman Group (2026)
ai-ux-patterns
-
OX Security 2026
62% of AI-built apps ship a critical vulndata-schema
-
Shape of AI
ai-ux-patterns
-
Shape Up
circuit-breaker / appetitefeature-flags
-
Shopify's modular monolith
scalable-architecture
-
Unleash / LaunchDarkly
trunk-based flags; flag-debt hygienefeature-flags
-
Veracode (Spring 2026)
insecure AI-generated codeagent-security testing-with-agents
Research, reports & incidents
Papers, studies and the named failures BOSS designs against.
-
CDT — Dark Patterns in AI Chatbots (2026)
ai-ux-patterns
-
CVE-2025-48757 (Lovable/Supabase RLS class)
data-schema ship-it-live
-
DORA / Accelerate
ROI of AI-Assisted Software Development (2026.01)git-workflow ship-it-live
-
How Coding Agents Fail Their Users (arXiv 2605.29442)
20,574-session misalignment analysistesting-with-agents
-
Professional Software Developers Don't Vibe, They Control (arXiv 2512.14012)
testing-with-agents
-
The MoltBook credential exposure
data-schema ship-it-live
-
Vaccaro et al., CHI'26
generated code injects dark patternsai-ux-patterns
Standards & specifications
The specs and canonical documentation BOSS reads directly.
-
Apple HIG — generative AI
ai-ux-patterns
-
MCP deprecated-features registrykey
mcp
-
MCP specification (2026-07-28)key
mcp
-
OWASP — Agentic ASI Top 10 / LLM Top 10key
agent-security ship-it-live
-
PostgreSQL row-level-security docskey
data-schema
-
W3C Design Tokens Community Groupkey
design-system
Crediting BOSS is optional, and it stays optional
BOSS's own promise is that it never puts its name on your work — you ship it, your name's on it. That's a promise about taking credit, not about refusing to let anyone give it. So nothing is ever added automatically, and nothing appears anywhere your users see.
If you want to say it anyway, boss credit shows you one line for your README and
--apply adds it. --remove takes it back out and leaves your own
words exactly as they were.
<!-- Builds, Or Stays Silent. ✦ --> Made with [BOSS](https://oyeboss.build) — *Build Out Solid Stuff.*
The comment is one of BOSS's alternate full forms — a wink for whoever reads the source, rather than a badge.
And one thing worth saying out loud: BOSS writes a marked
<!-- boss: --> comment into your CLAUDE.md so it can remove
itself cleanly later. That marker is also findable by searching public repositories, which is
honestly how anyone would know BOSS is being used at all. /welcome says so when
you start, and boss remove takes it out along with everything else. Counting
people who never agreed to be counted is surveillance, however public the data — so it's
disclosed rather than discovered.
Naming everyone is right. Linking everyone is busywork.
Everybody above is credited by name. Only the 20 marked key owe a link — a named practitioner whose thinking is distilled directly into a practice, or a primary spec a reviewer would re-open. A supporting citation sitting behind one claim doesn't need a URL to have been honest.
For the key ones it matters, because an attribution without a URL is half a
citation: it can't be cheaply re-checked when the practice comes up for review, the
person can't be credited properly, and the next reader redoes a search someone already did.
The verification work was already done — /vet goes to the primary source before
it grades anything — and then the URL was thrown away.
So the rule now sits inside /vet: write down the URL you actually opened —
not the thread that pointed at it. Each name becomes a link the moment its URL lands.
Currently 19 of 20 key sources linked.