“Prompt engineering”

Transform poor or overly simple prompts with expert-level framing.

Design, audit, run, or improve bounded agent loops whose feedback changes the next action.

Create a handoff brief for another LLM to implement.

Integrate systems after frontend direction.

Prove a system survives failure by injecting it deliberately — game days, fault injection, and steady-state hypotheses.

Compose installed agent skills into one orchestrated multi-agent run: an agent graph where each node is a sub-agent that loads exactly one skill, edges carry…

Run bounded, evidence-driven loops for repeated, resumable, delegated, or scheduled work.

Use for ingestion, ETL or ELT, data pipelines, mapping, cleansing, validation, reconciliation, lineage, warehousing, dimensional models, semantic layers, file…

Use when improving prompts for any LLM.

Production bundle combining baseline, issue, contract, deployment, data, configuration, observability, and security controls.

精通机器学习模型开发与部署的 AI 工程专家，擅长从数据处理到模型上线的全链路工程化，专注构建可靠、可扩展的 AI 系统。

Feature construction from market data for ML trading models including price, volume, on-chain, and microstructure features

Use when Malvex needs to turn validated findings into YARA, Sigma, EDR, network, memory, or behavioral detections.

Clarify vague or underspecified creative intent before prompt construction.

Apply general production engineering standards when implementing, fixing, reviewing, or refactoring code in any language or framework.

Select focused debugging, test-first or architecture guidance for code implementation, failures and tests.

Cache the parts of the prompt that don't change so a long-running loop stops paying full price on every turn.

Resilience engineering, fault injection testing, circuit breaker patterns, and graceful degradation protocols.

Expert application security engineer specializing in threat modeling, vulnerability assessment, secure code review, and security architecture design.

General implementation only when no narrower task or ecosystem skill fits.

Orchestrator for ad-hoc work and for a long chat.

Apply Graph Engineering workflows inside Hermes.

Engineering principles (inviolable) for the software-delivery fabric.

Use explicit human checkpoints for architecture, security, rollout, or operational decisions.