Patrick Herron
Community@pjherron · Durham, NC, US
Patrick Herron maintains a 72-skill registry for Claude Code covering agentic engineering, testing, database design, cloud deployment, and agent memory systems.
Agent Skills by Patrick Herron
Showing 72 vetted skills indexed across 1 GitHub repositories.
pjherronknowledge-ops
Organizes, syncs, and retrieves knowledge across memory files, MCP memory, Git repos, and external data stores.
pjherrondomain-modeling
Build and maintain project domain glossaries and architectural decision records.
pjherronsession-recruitment
Queries the OpenCode SQLite database to list, search, and resume past sessions.
pjherronagent-introspection-debugging
Diagnoses and recovers failing AI agent runs through structured capture, root-cause analysis, and introspection reports.
pjherroncode-tour
Creates CodeTour .tour files with persona-targeted walkthroughs anchored to real files and lines.
pjherronpython-testing
Implements pytest test suites with fixtures, mocking, parametrization, and coverage reporting.
pjherronsas-migration
Migrate SAS programs to Python, Java, or Postgres with operational fidelity verification.
pjherrondocker-patterns
Provides Docker and Docker Compose patterns for local development, networking, volumes, and container security.
pjherronskill-invoke
Presents an Accept, Review, Revise, or Decline consent prompt before any skill executes.
pjherronpostgres-patterns
Provides PostgreSQL patterns for indexing, schema design, query optimization, and row-level security.
pjherronDatabase Design
Design relational and document database schemas with normalization, indexing, and migration patterns.
pjherronHooks Automation
Automates pre/post-operation hooks for Claude Code with formatting, memory coordination, and neural pattern training.
pjherronenterprise-agent-ops
Operate long-lived agent workloads with observability, security boundaries, and lifecycle management.
pjherronbootstrap
Loads prior session context and recruits relevant skills at session start.
pjherronObservability & Monitoring
Implement structured logging, metrics, distributed tracing, and alerting for production systems.
pjherronskill-recruitment
Discovers relevant skills and past session patterns from OpenCode history and skill files.
pjherronVerification & Quality Assurance
Verify code quality with truth scoring, automated checks, and git-based rollback.
pjherroncoding-standards
Enforces baseline naming, immutability, and readability conventions across TypeScript and JavaScript projects.
pjherroneval-harness
Implements eval-driven development with pass/fail criteria and pass@k metrics for AI coding sessions.
pjherroncost-aware-llm-pipeline
Implement model routing, budget tracking, retry logic, and prompt caching for LLM API pipelines.
pjherronSkill Builder
Create Claude Code Skills with YAML frontmatter and progressive disclosure structure.
pjherronmodel-roster-review
Review and build LLM model rosters for agentic routing tiers following ADR 0007.
pjherrongit-workflow
Guides Git branching strategies, commit conventions, and collaborative development workflows.
pjherronshadow-profile
View and reset personal skill usage profiles stored in PostgreSQL.
Frequently Asked Questions About Patrick Herron
FAQPage SchemaWhat tasks can I accomplish with pjherron's skill registry?▼
The 72 skills cover agentic engineering, TDD with pytest and Playwright, PostgreSQL schema design and migrations, Docker/Kubernetes/AWS deployment, GitHub issue and PR operations, SAS-to-Python migration, API contract testing, security scanning, and building self-learning agents with AgentDB vector memory.
Who is the target audience for these skills?▼
Software engineers and platform teams running Claude Code who need structured patterns for backend (FastAPI, Node.js, Next.js), frontend (React), database design, CI/CD deployment, and agent operations. Planning skills like grill-me, to-spec, and to-tickets target tech leads decomposing features into tickets.
How are these skills installed and invoked?▼
The configure-ecc skill provides an interactive installer that places skills into user-level or project-level directories and verifies paths. The bootstrap skill loads prior context at session start, while skill-invoke wraps invocation with an Accept/Review/Revise/Decline consent step.
What prerequisites or dependencies do these skills require?▼
Skills assume a Claude Code environment with MCP server support. Specific skills declare dependencies: github-code-review requires github-cli, ruv-swarm, and claude-flow; deep-research uses firecrawl and exa MCPs; eval-harness and skill-comply declare Read, Write, Edit, Bash, Grep, and Glob tool access.
How do the skills manage cost and context consumption?▼
cost-aware-llm-pipeline routes models by task complexity with budget tracking and prompt caching. token-budget-advisor lets users control response depth, context-budget audits token consumption across agents and MCP servers, and strategic-compact suggests manual compaction at logical task boundaries.