Zhachory Volker
Community@zhachory1 · New York, NY
Pretty swell guy. 12+ years of SWE and ML experience. I care about responsible AI, computational journalism, and the systems helping organize the knowledge.
Agent Skills by Zhachory Volker
Showing 18 vetted skills indexed across 1 GitHub repositories.
Zhachory1run-telemetry
Emit structured JSONL telemetry events for AI agent workflow runs.
Zhachory1grounding-brief
Gathers evidence from memory, code search, logs, and tickets into a structured context brief.
Zhachory1project-ideation
Transforms goals and OKRs into evidence-backed, feasibility-checked project candidates for roadmap evaluation.
Zhachory1handoff-packaging
Package workflow outputs into structured handoff bundles with evidence, decisions, and next actions.
Zhachory1roadmapping
Sequences project candidates into a prioritized roadmap with ROI scoring, dependencies, and capacity allocation.
Zhachory1task-decomposition-planning
Convert approved goals and designs into ordered, dependency-aware tasks with acceptance criteria.
Zhachory1plan-to-launch
Orchestrates end-to-end software development from PRD and design docs through implementation, review, and launch approval.
Zhachory1human-approval-gate
Prepares decision packages with evidence, risks, and rollback plans for human signoff.
Zhachory1success-criteria-metrics
Define measurable success criteria, guardrails, and decision thresholds before work begins.
Zhachory1backprop
Optimizes AI agent workflows by analyzing run telemetry, failure clusters, and A/B rollout results.
Zhachory1adversarial-probe
Probes services and AI-agent workflows with load, chaos, fuzzing, and adversarial inputs to surface breakages.
Zhachory1debug-investigation
Investigate bugs and incidents through evidence gathering, hypothesis testing, and root cause analysis.
Zhachory1hypothesis-testing
Generates and tests falsifiable hypotheses for ML experiments and debugging loops.
Zhachory1ml-experiments
Runs ML experiments from metric definition through reproducible artifact handoff.
Zhachory1structured-doc-authoring
Author structured PRDs, design docs, plans, ADRs, roadmaps, and RCAs with templates and diagrams.
Zhachory1pr-review
Reviews pull requests against intent and publishes verdict-mapped GitHub reviews with prioritized findings.
Zhachory1council
Convene 3-6 specialist personas to debate and review high-stakes decisions.
Zhachory1ship
Implements accepted specs through a scoped patch, test, docs, and validation workflow.
Frequently Asked Questions About Zhachory Volker
FAQPage SchemaWhat tasks can I accomplish with Zhachory1's skills?▼
You can run end-to-end agent engineering: gather grounding context, decompose goals into dependency-aware tasks, author PRDs and design docs, execute scoped code patches via ship, run multi-persona council reviews, debug incidents, run ML experiments, and package handoffs with telemetry.
Who are these skills designed for?▼
They target software engineers, ML practitioners, and agent-fleet operators who need disciplined, evidence-driven processes. Personas include engineers shipping PRs, ML researchers running tracked experiments, and leads needing decision-grade council reviews and human approval gates.
How do the skills enforce safety and accountability?▼
The human-approval-gate skill prepares decision packages with recommendations, evidence, risks, and rollback plans, and never auto-approves accountable decisions. Adversarial-probe requires locked breakage thresholds and mandatory approval before destructive or live probes.
What telemetry and optimization capabilities are included?▼
Run-telemetry emits structured JSONL events capturing latency, cost, tokens, failure rate, human-edit rate, and PR outcomes. The backprop skill analyzes run history to drive prompt and skill improvements, A/B tests, and promote-or-rollback changelogs.
What prerequisites or dependencies do these skills assume?▼
Skills reference long-term memory MCP, code RAG, repo-index, and an agent-fleet journal for context. PR review defaults to publishing GitHub reviews, so repository access and a GitHub-integrated environment are assumed for full functionality.