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
Use when designing or tuning intake controls for PR volume, merge-queue admission, and reviewer/runner capacity limits. USE FOR: WIP limit policies, queue admission rules, stale PR handling, and safe load-shedding design. DO NOT USE FOR: overriding branch protections manually, skipping approvals, or implementing…
Agent Claude Code
Use when auditing PR intake, merge queue health, CI tier efficiency, and stale-branch pressure across repositories. USE FOR: queue bottleneck audits, CI waste analysis, PR aging reviews, merge-path risk scoring, and evidence-backed findings. DO NOT USE FOR: writing feature code, merging PRs without approval, or…
Agent Claude Code
Use when coordinating flow-governance analysis across code, PRs, issues, skills, and scripts. USE FOR: reusing existing flow/CI/governance skills first, mapping findings to current agents and scripts, identifying missing skill/script coverage, and producing issue-ready gap backlogs. DO NOT USE FOR: writing product…
Agent Claude Code
Use when building an optimization program that combines audit evidence with prioritized fixes and execution sequencing. USE FOR: phased flow optimization plans, merge-queue and CI policy tuning, target KPI definition, and rollout risk reduction. DO NOT USE FOR: one-off bugfix coding, ad hoc release approvals, or…
Agent Claude Code
Use when converting audit findings into prioritized, issue-ready fixes with acceptance criteria and owners. USE FOR: translating audit evidence into GitHub issues, prioritizing queue and CI fixes, auto-creating high-confidence issues, sequencing remediation work, and defining measurable outcomes. DO NOT USE FOR…
Agent Claude Code
Use when tracking repository flow metrics over time and recommending corrective actions when trends degrade. USE FOR: weekly flow scorecards, trend analysis, regression alerts, and metric-backed improvement suggestions. DO NOT USE FOR: speculative architecture redesign, editing production code, or replacing incident…
Agent Claude Code
Use when auditing issue and PR metadata, label drift, template gaps, and workflow enforcement coverage. USE FOR: label audits, template consistency checks, workflow drift reviews, gap identification, and issue-ready findings. DO NOT USE FOR: writing application code, implementing product features, or changing labels…
Agent Claude Code
Use when writing or revising common-vs-specific governance guidance, canonical label contracts, and migration notes. USE FOR: drafting docs/reference/governance-contract.md, separating shared rules from repo-specific rules, mapping legacy labels to canonical labels, and planning governance issue backlogs. DO NOT USE…
Agent Claude Code
Compatibility alias for the orchestrator agent. Preserves the legacy filename while the prefixed BaseCoat agent is the canonical source. USE FOR: routing multi-step workflows, coordinating parallel agents, tracking execution state. DO NOT USE FOR: single-step tasks, user-facing triage, direct tool calls.
Agent Claude Code
Thin orchestration entrypoint for end-to-end startup pack generation. USE FOR: bootstrapping a new program with coordinated onboarding/backlog/spec/architecture/workflow outputs, dry-run orchestration before writing artifacts, resuming partial orchestration with checkpoints, preserving repo-specific delivery labels…
Agent Claude Code
Runs a pull request-scoped triage loop for inbound automated Copilot review threads. USE FOR: impact-sizing review threads, validating Medium suggestions, recording safe dispositions, resolving Small and Medium threads, and escalating Large findings. DO NOT USE FOR: initial pull request review, auto-applying security…
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: