Menu
Skills in this repository
SkillsMP has collected 311 skills from lukemcqueen/hermes-cortex. Open a skill to review its source and details.
lukemcqueen/hermes-cortexShowing 31 of 311 collected skills.
Systematic "system is slow" diagnosis — baseline resource check, CPU frequency scaling analysis, process/core affinity mapping, container-level activity, and differentiation between transient vs chronic bottlenecks.
Systematic Linux server hardening with tiered prioritization. Covers UFW firewall, SSH hardening (key-only auth, timeouts, rate limits), unattended-upgrades, kernel sysctl tuning, fail2ban jail expansion, and audit tooling. Uses a three-tier framework (high…
A warm, practical Korean language companion for English speakers aged 50+. Covers Hangul mastery, essential grammar, honorifics, pronunciation, Anki strategies for older learners, reading progression, cultural context, grammar reference tables, and real-world…
Source text: Mixed languages
Use when checking if a software/tool name is available for use — searches GitHub, web, and registries for conflicts, evaluates severity, and generates alternatives.
Interact with FreshTomato/DD-WRT routers programmatically via curl — authentication, nvram access, port forwarding, DMZ, and configuration management.
Next.js Docker multi-stage builds with standalone output — minimal runtime images, no node_modules in production
Handle passwords, API keys, tokens, and secrets securely when using terminal/read_file/execute_code tools — avoid exposing sensitive values in tool call parameters or chat output.
Two-axis pre-commit review: Standards (documents + code smells) and Spec (requirement compliance) via parallel sub-agents. Plus security scan, quality gates, auto-fix.
Deep module vocabulary and design principles — module, interface, depth, seam, adapter, leverage, locality. Use when designing or improving a module's interface, finding deepening opportunities, deciding where a seam goes, or making code more testable.
Error handling patterns and idioms: structured exceptions, graceful degradation, retry strategies, circuit breakers, user-facing error messages, and failure domain boundaries across Python and web apps.
Universal lesson-aware injection pattern. Makes every agent action memory-aware: search lessons before acting, save lessons after fixing. Works across all skills, not just change-test-loop.
Structured logging conventions: log levels, format standards, context injection, correlation IDs, sensitive data scrubbing, and rotation. Applies to Python, Node.js, shell scripts, and infrastructure.
After any improvement, bug fix, workflow discovery, or lesson — pause and evaluate whether the insight is public-worthy for the Hermes Cortex OSS community. If yes, genericize and contribute.
70+ React & Next.js performance optimization rules from Vercel Engineering — covers waterfalls, bundle size, server/client perf, re-renders, and JS micro-optimizations. Use when writing new components, reviewing code for perf issues, or optimizing data…
React composition patterns that scale — compound components, state lifting, context interfaces, and avoiding boolean prop proliferation. Includes React 19 API changes. Use when refactoring components with many boolean props or building flexible component…
Implement smooth native-browser animations between UI states using React's ViewTransition component and document.startViewTransition API. No third-party animation libraries needed. Covers shared element morphs, enter/exit, list reorder, and Next.js App Router…
6-phase root cause debugging: feedback loop, reproduce, pattern, hypothesise + instrument, fix, cleanup. Understand bugs before fixing.
Execute plans via delegate_task subagents (2-stage review).
Ensure test seed data never causes unique constraint violations — UUID-based, timestamp-based, counter-based, or Faker patterns for any field with a UNIQUE index, primary key, or slug constraint.
Enforce the "prove existing can't handle it" discipline before creating any new file. Supplements survey-before-action with the HARD RULE that every agent defaults to "extend existing" first. Prevents codebase fragmentation from unnecessary parallel systems.
Comprehensive integration audit — runs all changed subsystems together before final commit to catch cross-system gaps that individual testing misses. Covers registry validation, schema checks, cheat detection, deployed-state drift, doctor, and syntax…
CI/CD pipeline configuration patterns: GitHub Actions, multi-stage builds, testing matrices, deployment workflows, secrets management, caching strategies, and pipeline optimization.
Age-gated package installation protection. Before installing any package with pip, npm, brew, or cargo, verify it's been public for at least 14 days. Blocks supply-chain attacks on brand-new releases (dependency confusion, account takeover, malware).
Write clear, structured git commit messages following Conventional Commits format. Includes type prefixes, scope, breaking change markers, and body conventions for automated changelog generation.
Full PR review pipeline — whole-repo context, architecture analysis, lesson-DB pattern matching, test regression check, and formal review submission with inline comments. Zero external API costs. Replaces Greptile / PR-Agent with Hermes-native tooling.
API documentation standards and tooling: OpenAPI/Swagger specs, endpoint descriptions, request/response schemas, error documentation, authentication docs, changelogs, and generation tooling.
Plan mode: write an actionable markdown plan to .hermes/plans/, no execution. Bite-sized tasks, exact paths, complete code.
Concise 1-page PRD template: problem, scope, functional/non-functional requirements, edge cases, acceptance criteria, risks, rollout strategy, and success metrics.
Structural project analysis — build a dependency graph so agents don't guess which files a change affects. Trigger: before multi-file changes, entering a new project, code review. Run: project-map analyze Read: .hermes-cortex/project-map.json
Information routing decision matrix for Hermes Cortex agents. Defines when to consult live context vs session history vs memory vs docs, with explicit priority ordering, staleness handling, confidence scoring, and cross-reference chaining patterns for…
Local semantic web cache for AI agents — cache web_search and web_extract results locally with embeddings, reducing API costs and enabling offline operation.