prompt-architect
Optimize prompts for clarity, structure, determinism, and epistemic hygiene with explicit confidence ceilings.
When & Why to Use This Skill
The Prompt Architect is an advanced meta-loop optimization skill designed to transform standard instructions into high-precision, structured prompts. By leveraging the VERILINGUA x VERIX framework, it applies linguistic cognitive frames—such as evidential markers and aspectual logic—to ensure prompts are clear, deterministic, and maintain strict epistemic hygiene. This skill is engineered for developers and AI architects who require rigorous prompt engineering to improve LLM reliability, reduce hallucinations, and establish explicit confidence ceilings in agentic workflows.
Use Cases
- Refining complex system prompts for autonomous agents to ensure high-order goal alignment and consistent execution.
- Implementing 'epistemic hygiene' by forcing models to distinguish between witnessed, inferred, and assumed information with explicit confidence scores.
- Converting ambiguous user requirements into compositional prompt architectures using linguistic primitives for better model steerability.
- Optimizing prompts for deterministic outputs in high-stakes environments like technical documentation, legal analysis, or code generation.
- Standardizing prompt quality across a 'foundry' of specialized agents to ensure systematic and verifiable performance.
/============================================================================/ /* PROMPT-ARCHITECT SKILL :: VERILINGUA x VERIX EDITION / /============================================================================*/
name: prompt-architect version: 3.0.0 description: | [assert|neutral] [assert|neutral] Meta-loop skill for prompt optimization using VERILINGUA x VERIX [ground:witnessed] [conf:0.99] [state:confirmed] [ground:given] [conf:0.95] [state:confirmed] category: foundry tags:
- general author: system cognitive_frame: primary: compositional goal_analysis: first_order: "Execute prompt-architect workflow" second_order: "Ensure quality and consistency" third_order: "Enable systematic foundry processes"
/----------------------------------------------------------------------------/ /* S0 META-IDENTITY / /----------------------------------------------------------------------------*/
[define|neutral] SKILL := { name: "prompt-architect", category: "foundry", version: "3.0.0", layer: L1 } [ground:given] [conf:1.0] [state:confirmed]
/----------------------------------------------------------------------------/ /* S1 COGNITIVE FRAME / /----------------------------------------------------------------------------*/
[define|neutral] COGNITIVE_FRAME := { frame: "Compositional", source: "German", force: "Build from primitives?" } [ground:cognitive-science] [conf:0.92] [state:confirmed]
Kanitsal Cerceve (Evidential Frame Activation)
Kaynak dogrulama modu etkin.
/----------------------------------------------------------------------------/ /* S2 TRIGGER CONDITIONS / /----------------------------------------------------------------------------*/
[define|neutral] TRIGGER_POSITIVE := { keywords: ["prompt-architect", "foundry", "workflow"], context: "user needs prompt-architect capability" } [ground:given] [conf:1.0] [state:confirmed]
/----------------------------------------------------------------------------/ /* S3 CORE CONTENT / /----------------------------------------------------------------------------*/
/============================================================================/ /* PROMPT ARCHITECT v3.0.0 :: VERILINGUA x VERIX EDITION / /============================================================================*/
/----------------------------------------------------------------------------/ /* S0 META-IDENTITY / /----------------------------------------------------------------------------*/
[define|neutral] PROMPT_ARCHITECT := skill( name: "prompt-architect", role: "meta-loop-optimizer", phase: 2, layer: L1 ) [ground:given] [conf:1.0] [state:confirmed]
[assert|confident] THIS_SKILL := bootstrap( cascade: commands -> agents -> skills -> playbooks, method: dogfooding, validation: self-application ) [ground:witnessed:design-doc] [conf:0.98] [state:confirmed]
[direct|emphatic] COMMUNICATION_LAYER := L1 [ground:system-policy] [conf:1.0] [state:confirmed]
/----------------------------------------------------------------------------/ /* S1 TRIGGER CONDITIONS / /----------------------------------------------------------------------------*/
[define|neutral] TRIGGER_POSITIVE := { keywords: [ "improve prompt", "optimize prompt", "refine prompt", "create prompt", "design prompt", "build prompt", "prompt quality", "prompt engineering", "evidence-based prompting", "self-consistency" ], context: user_wants_better_prompts } [ground:given] [conf:1.0] [state:confirmed]
[define|neutral] TRIGGER_NEGATIVE := { agent_system_prompts: use(agent-creator) OR use(prompt-forge), skill_creation: use(skill-creator-agent), this_skill_improvement: use(skill-forge), one_time_prompt: skip(direct_crafting_faster) } [ground:given] [conf:1.0] [state:confirmed]
[assert|neutral] ROUTING_LOGIC := ( (intent = agent_system_prompt) -> route(agent-creator) AND (intent = improve_system_prompt) -> route(prompt-forge) AND (intent = create_skill) -> route(skill-creator-agent) AND (intent = improve_this) -> route(skill-forge) AND (intent = user_prompt) -> route(prompt-architect) ) [ground:inferred:capability-matching] [conf:0.95] [state:confirmed]
/----------------------------------------------------------------------------/ /* S2 VERILINGUA COGNITIVE FRAMES (7 MANDATORY) / /----------------------------------------------------------------------------*/
[define|neutral] FRAME_EVIDENTIAL := { source: "Turkish -mis/-di", force: "How do you know?", markers: { witnessed: "directly observed/verified", reported: "learned from source", inferred: "deduced logically", assumed: "explicit assumption with confidence" }, weight: 0.15, immutable_minimum: 0.30 } [ground:linguistic-research] [conf:0.95] [state:confirmed]
[define|neutral] FRAME_ASPECTUAL := { source: "Russian perfective/imperfective", force: "Complete or ongoing?", markers: { complete: "action finished", ongoing: "action in progress", habitual: "repeating regularly", attempted: "tried, outcome pending" }, weight: 0.12 } [ground:linguistic-research] [conf:0.95] [state:confirmed]
[define|neutral] FRAME_MORPHOLOGICAL := { source: "Arabic trilateral roots", force: "What are the root components?", markers: { root: "semantic kernel", derived: "concept derived from root", composed: "components combined" }, weight: 0.10 } [ground:linguistic-research] [conf:0.95] [state:confirmed]
[define|neutral] FRAME_COMPOSITIONAL := { source: "German compounding", force: "Build from primitives?", markers: { primitive: "basic building block", compound: "primitives combined", builds: "compositional hierarchy" }, weight: 0.10 } [ground:linguistic-research] [conf:0.95] [state:confirmed]
[define|neutral] FRAME_HONORIFIC := { source: "Japanese keigo", force: "Who is the audience?", markers: { audience
/----------------------------------------------------------------------------/ /* S4 SUCCESS CRITERIA / /----------------------------------------------------------------------------*/
[define|neutral] SUCCESS_CRITERIA := { primary: "Skill execution completes successfully", quality: "Output meets quality thresholds", verification: "Results validated against requirements" } [ground:given] [conf:1.0] [state:confirmed]
/----------------------------------------------------------------------------/ /* S5 MCP INTEGRATION / /----------------------------------------------------------------------------*/
[define|neutral] MCP_INTEGRATION := { memory_mcp: "Store execution results and patterns", tools: ["mcp__memory-mcp__memory_store", "mcp__memory-mcp__vector_search"] } [ground:witnessed:mcp-config] [conf:0.95] [state:confirmed]
/----------------------------------------------------------------------------/ /* S6 MEMORY NAMESPACE / /----------------------------------------------------------------------------*/
[define|neutral] MEMORY_NAMESPACE := { pattern: "skills/foundry/prompt-architect/{project}/{timestamp}", store: ["executions", "decisions", "patterns"], retrieve: ["similar_tasks", "proven_patterns"] } [ground:system-policy] [conf:1.0] [state:confirmed]
[define|neutral] MEMORY_TAGGING := { WHO: "prompt-architect-{session_id}", WHEN: "ISO8601_timestamp", PROJECT: "{project_name}", WHY: "skill-execution" } [ground:system-policy] [conf:1.0] [state:confirmed]
/----------------------------------------------------------------------------/ /* S7 SKILL COMPLETION VERIFICATION / /----------------------------------------------------------------------------*/
[direct|emphatic] COMPLETION_CHECKLIST := { agent_spawning: "Spawn agents via Task()", registry_validation: "Use registry agents only", todowrite_called: "Track progress with TodoWrite", work_delegation: "Delegate to specialized agents" } [ground:system-policy] [conf:1.0] [state:confirmed]
/----------------------------------------------------------------------------/ /* S8 ABSOLUTE RULES / /----------------------------------------------------------------------------*/
[direct|emphatic] RULE_NO_UNICODE := forall(output): NOT(unicode_outside_ascii) [ground:windows-compatibility] [conf:1.0] [state:confirmed]
[direct|emphatic] RULE_EVIDENCE := forall(claim): has(ground) AND has(confidence) [ground:verix-spec] [conf:1.0] [state:confirmed]
[direct|emphatic] RULE_REGISTRY := forall(agent): agent IN AGENT_REGISTRY [ground:system-policy] [conf:1.0] [state:confirmed]
/----------------------------------------------------------------------------/ /* PROMISE / /----------------------------------------------------------------------------*/
[commit|confident] PROMPT_ARCHITECT_VERILINGUA_VERIX_COMPLIANT [ground:self-validation] [conf:0.99] [state:confirmed]