Enterprise shared library of GitHub Copilot customization assets — agents, skills, instructions, and prompts.
This repository also configures its own agents. See what basecoat tells them →
Agent Claude Code
Audit, measure, implement, and operate rules-based human-in-the-loop autonomy for agent-operated repositories. USE FOR: PR risk classification, auto-merge policy, merge queue gates, deployment lane policy, and policy-versus-settings drift. DO NOT USE FOR: product feature implementation, routine code review, or…
Agent Claude Code
Agent factory specialist. USE FOR: creating agent specs, auditing agent/skill definitions, capability-based routing decisions, and revising weak specs. DO NOT USE FOR: product feature coding, infrastructure deployment, unrelated code review.
Agent Claude Code
Agent operations and infrastructure specialist. USE FOR: monitoring agent health, tuning agent performance, debugging agent failures, optimizing resource usage. DO NOT USE FOR: individual agent tasks, direct coding.
Agent Claude Code
REST API and contract design specialist. USE FOR: designing RESTful APIs, creating OpenAPI specifications, planning API versioning strategies. DO NOT USE FOR: API implementation, testing.
Agent Claude Code
Scans legacy applications to discover dependencies, identify technology stacks, assess migration complexity, and generate architecture diagrams for portfolio analysis. USE FOR: scan legacy app tech stack, assess migration complexity, generate dependency inventory. DO NOT USE FOR: writing new code, live production…
Agent Claude Code
Backend and services development specialist. USE FOR: implementing backend services, designing server architectures, managing databases. DO NOT USE FOR: frontend work, DevOps operations.
Agent Claude Code
Use when validating a Workcell BOM against the plant registry before S2 starts. USE FOR: schema validation, CAF naming checks, circular dependency detection, and GitHub check results. DO NOT USE FOR: runtime state merges or cutover decisions.
Agent Claude Code
Use when cleaning stale branches, dangling refs, and release branch hygiene while preserving active work. USE FOR: identify merged or stale branches, prune safe remote refs, flag release branch drift, and produce cleanup actions with owners and due dates. DO NOT USE FOR: deleting branches with open PRs, changing…
Agent Claude Code
Designs layered CI/CD isolation so independent domains can evolve and release separately without cross-triggered pipeline noise. USE FOR: path-based pipeline routing, layered CI/CD boundaries, independent release lanes, cross-trigger prevention, and domain ownership isolation. DO NOT USE FOR: application feature…
Agent Claude Code
Chaos engineering agent for fault injection, game days, resilience scoring, recovery validation, and SLO-aware resilience experiments. USE FOR: design fault injection experiments, run game day exercises, validate SLO resilience targets. DO NOT USE FOR: debugging application bugs, writing unit tests.
Agent Claude Code
Contract Testing Agent for consumer-driven contracts, E2E testing strategy, and mutation testing for distributed systems. USE FOR: write Pact consumer-driven contracts, design service integration tests, set up contract verification in CI. DO NOT USE FOR: load testing, manual QA workflows.
Agent Claude Code
Validate that a feature, PR, or release meets the Definition of Done before closing. Enforces testing evidence, config verification, response validation, and acceptance criteria. USE FOR: check PR meets DoD, validate acceptance criteria, verify release readiness. DO NOT USE FOR: writing acceptance criteria…
Agent Claude Code
Use when you need to detect delivery gaps, cluster issues/PRs by function, map dependencies, and operationalize execution in a GitHub Project. USE FOR: identifying issue/PR patterns, grouping backlog by feature/function, mapping hard vs soft dependencies, creating/updating a GitHub Project and linking items. DO NOT…
Agent Claude Code
Triage and prioritize open Dependabot PR queues by CVE severity, semver risk, ecosystem impact, and blocking dependency chains. USE FOR: rank Dependabot PRs for merge order, identify security-blocking updates, group safe batch merges, surface dependency chain blockers. DO NOT USE FOR: creating Dependabot config…
Agent Claude Code
Use when monitoring workcell BOM dependencies and cell health to open or resolve blocker issues when a cell fails or recovers. USE FOR: read workcell BOM dependencies, monitor cell deployment status, create blocker issues, and resolve them on recovery. DO NOT USE FOR: general incident response, application code, or…
Agent Claude Code
Agent for managing dependency updates, tracking breaking changes, planning upgrade paths, monitoring vulnerabilities, analyzing semantic versioning, and generating migration guides. USE FOR: plan upgrade paths, generate migration guides, audit dependency CVEs. DO NOT USE FOR: writing new features, infrastructure…
Agent Claude Code
Reviews Dependabot pull requests and posts a structured risk assessment comment: semver bump type, breaking change likelihood, test focus, and CVE context. USE FOR: assess Dependabot PR risk, semver bump safety, CVE context. DO NOT USE FOR: creating Dependabot config, general code review.
Agent Claude Code
DevOps and infrastructure automation specialist. USE FOR: designing CI/CD pipelines, managing infrastructure-as-code, optimizing deployment processes. DO NOT USE FOR: application code, product feature development.
Agent Claude Code
Use when high-risk decisions need a human approver and a PR-comment approval trail. USE FOR: release signoff, irreversible changes, compliance gates. DO NOT USE FOR: routine automation.
Agent Claude Code
Generates mission-driven charters with scope, triage routing, and evidence capture for time-boxed exploratory testing. USE FOR: exploratory test charters, risk-based test missions, evidence capture, defect triage, and automation candidate discovery. DO NOT USE FOR: automated regression implementation, unit-test…
Agent Claude Code
Use when reading .factory-state.json and auto-queuing the next ready wave of factory work. USE FOR: route S2/S3/S4 work, batch dispatch by wave, and ping the right follow-up workflow. DO NOT USE FOR: state normalization, BOM validation, or product backlog prioritization.
Agent Claude Code
Use when merging Workcell intake YAML, GitHub labels, and gate results into a single S1-S5 state snapshot. USE FOR: normalize station state, publish .factory-state.json, reconcile blockers, and surface stale work. DO NOT USE FOR: implementing product code or changing workflow policy.
Agent Claude Code
Continuous learning and optimization through user feedback collection, prompt effectiveness tracking, outcome measurement, A/B testing, regression detection, and instruction refinement. USE FOR: track prompt effectiveness, run A/B test on instructions, detect quality regression. DO NOT USE FOR: product feature…
Agent Claude Code
FinOps advisor for cloud cost governance, cost optimization, chargeback/showback models, and 12-Factor App best practices for cost efficiency. USE FOR: analyze cloud spend by service, build chargeback/showback models, identify cost optimization opportunities. DO NOT USE FOR: live incident response, infrastructure…
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: