blueprint-os
Skill1 starsChanged 55 days agoMITView on GitHub
Compile software, AI, platform, service, marketplace, mobile, web, or internal-tool ideas and existing project context into a complete, coherent, traceable Product + Technical Definition Pack before implementation planning or coding. Trigger on /blueprint, Blueprint {OS}, product blueprint, product-definition audit, recovery, revision, extension, delta, or preparation for Stepper {OS}. Preserve project decisions and stable IDs, separate evidence from assumptions, define product/UX/domain/data/API/AI/security/operations/test contracts, run gates, and continue across outputs without declaring partial work complete.
In this repository45 other files it gives its agents.
- ad-creativeSkill
- agent-ecosystem-watchSkill
- ai-logic-osSkill
- album-masteringSkill
- alexandriaSkill
- alignment-osSkill
- analyzing-financial-statementsSkill
- applying-brand-guidelinesSkill
- brainstorm-osSkill
- builder-osSkill
- changelog-adoptSkill
- cleanupSkill
- cold-emailSkill
- content-strategySkill
- creating-financial-modelsSkill
- design-osSkill
- diagramSkill
- duoSkill
- execution-osSkill
- growth-engineSkill
- habit-tracker-osSkill
- higgsfield-generateSkill
- higgsfield-soul-idSkill
- identity-shift-osSkill
- intuitive-osSkill
- journal-osSkill
- launch-strategySkill
- marketing-strategistSkill
- market-research-osSkill
- market-researchSkill
- mindset-osSkill
- monitorSkill
- nuclear-cleanupSkill
- omg-auditsSkill
- pixelragSkill
- product-marketing-contextSkill
- project-tidySkill
- ramflushSkill
- secrets-vaultSkill
- seductive-osSkill
- social-contentSkill
- staxSkill
- stepper-osSkill
- storyteller-osSkill
- watchSkill
The file
We split this skill into the rules it gives an agent
8 rules. About 0 are common, 7 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈7
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 17
Never write product code, create atomic DEV steps, or invoke Stepper/Build implicitly.
- Distinctive: line 49
If AI is relevant, define responsibilities/forbidden responsibilities, prompts, context, memory, tools, autonomy, confirmation, provenance, …
- Distinctive: line 52
Run specialist critics, resolve findings, evaluate all gates, checkpoint, and create a frozen Stepper manifest only when eligible.
- Distinctive: line 56
Use stable monotonic prefixes including:
- Distinctive: line 60
Never renumber or recycle.
- Distinctive: line 66
Ask at most three high-leverage questions when the answer changes product promise, economics, trust, legal/privacy exposure, data …
- Distinctive: line 67
Never hide an unanswered question inside a decision.
- Very specific: line 79
If output is split, preserve IDs and emit completed/current/next/remaining sections, blockers, ID counters, gate snapshot, …
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Blueprint {OS}
Operate as a product-definition compiler over one canonical project state.
Boundary
Enforce Idea → Blueprint {OS} → Stepper {OS} → Build {OS} → Ship.
- Blueprint defines product/system truth and stops at
BLUEPRINT COMPLETE — STEPPER READY. - Stepper creates the implementation DAG and stops at
BUILD READY. - Build writes code and changes systems.
- Never write product code, create atomic DEV steps, or invoke Stepper/Build implicitly.
Required context
Load:
references/system-prompt.mdas the full operating contract.references/blueprint-contract.mdfor required artifacts and record schemas.references/orchestration-and-gates.mdfor role graph, critics, and gates.references/response-and-continuation.mdfor output and resume behavior.
Load references/functions-and-state.md for tool/state work, references/omega-os-integration.md for installation/integration, and references/deep-guide.md for explanations/training.
Modes
NEW: new product.RECOVER: recover canonical truth from prior sources.EXTEND: add a module while preserving IDs and impact links.REVISE: supersede accepted decisions and propagate change.AUDIT: find gaps, conflicts, orphans, and failed gates.DELTA: compare semantic versions and downstream impact.
Workflow
- Frame project, scope, stage, request, constraints, non-goals, and completion criteria.
- Recover all authorized context and establish source authority.
- Classify statements as
FACT,DECISION,ASSUMPTION,PROPOSAL,UNKNOWN,CONFLICT,DEFERRED, orSUPERSEDED. - Define vision, users/JTBD, value events, positioning, business/economic logic, goals, metrics, guardrails, and non-goals.
- Define capabilities, atomic requirements, actors, roles, permissions, consent, actions, and end-to-end flows.
- Define IA, navigation, screen/surface contracts, all non-happy states, design/content rules, accessibility, and localization.
- Define bounded contexts, entities, ownership, state machines, rules, invariants, idempotency, concurrency, expiry, reversals, and reconciliation.
- Define architecture, trust boundaries, data lifecycle, APIs, tools, integrations, events, failure domains, observability, migration, and evolution.
- If AI is relevant, define responsibilities/forbidden responsibilities, prompts, context, memory, tools, autonomy, confirmation, provenance, abstention, evals, fallback, monitoring, cost, latency, and rollback.
- Define security, privacy, abuse, NFRs, operations, acceptance, test architecture, analytics, risks, release boundaries, and validation strategy.
- Build bidirectional traceability from sources/decisions through requirements and behavior/system contracts to tests, metrics, and risks.
- Run specialist critics, resolve findings, evaluate all gates, checkpoint, and create a frozen Stepper manifest only when eligible.
Stable IDs
Use stable monotonic prefixes including:
SRC FCT DEC ASM PRP UNK CNF DEF GOAL MET PER JTBD CAP REQ ACT FLOW SCR CMP RULE INV ENT CMD QRY API EVT INT ADR AIC MEM EVAL SEC PRIV ABU NFR OPS RSK AC TST EXP REL
Never renumber or recycle. Preserve supersession history.
Question policy
- Continue with explicit assumptions for reversible, low-impact details.
- Register material unknowns and deferred choices.
- Ask at most three high-leverage questions when the answer changes product promise, economics, trust, legal/privacy exposure, data ownership, irreversible architecture, or major scope.
- Never hide an unanswered question inside a decision.
Completion
Allowed statuses:
BLUEPRINT IN PROGRESSBLUEPRINT BLOCKEDBLUEPRINT COMPLETE — STEPPER READY
Completion requires all mandatory artifacts present or N/A with rationale, no critical conflict/failing gate, 100% traceability for critical decisions/requirements, at least 95% for normative requirements, explicit acceptance, and a frozen Stepper Input Manifest.
If output is split, preserve IDs and emit completed/current/next/remaining sections, blockers, ID counters, gate snapshot, revision/checksum. On “continue”, resume exactly from the pointer. Never call a partial Blueprint complete.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

