A Claude Code subagent that helps you pick the right Anthropic Claude model in GitHub Copilot's model picker and decide when to switch to Claude Code instead.
Open the source and read safety notes before installing.
Source-backed facts for citing this resource, derived directly from the registry — also available as plain text for AI assistants.
Decision playbook
Signals are present but mixed. Use the checklist below to confirm the source and operational safety for your environment.
0
78
—
No baseline selected
No major trust-signal divergence detected in the current selection.
Confirm ownership and provenance before trusting install instructions.
Source link availableRequired
Open the canonical repository and verify ownership.
Source provenance statusRequired
Marked as source-backed.
Metadata reviewed
Registry metadata indicates a reviewed listing.
Validate risk disclosures before installation or API wiring.
Safety notes presentRequired
Review the listed safety guidance before running commands.
Privacy notes presentRequired
Review data handling notes before connecting accounts or secrets.
Trust level risk gateRequired
Trust level does not block evaluation.
Check package metadata and artifact integrity signals.
Install payload available
Install or copy payload is available for review.
Package verification flag
No package verification flag provided.
Checksum metadata
No checksum provided for downloaded artifact.
Use compare context to validate trade-offs before adoption.
Compare tray has multiple entries
Add at least one more entry to compare trust differences.
Baseline comparison available
No baseline peer selected yet.
Diverging trust signals identified
No major trust-signal divergence found.
Setup at a glance
Copy-ready — paste the snippet to get started.
Adoption plan
Current risk score 16/100. Use staged verification before broader rollout.
Validate source and review signals before any execution.
Confirm source provenanceRequired
Source URL/provenance metadata is present.
Confirm metadata review state
Listing has review metadata.
Verify install payload
Install/config payload exists and can be inspected.
Confirm safety, privacy, and package integrity signals.
Review safety notesRequired
Safety notes are present.
Review privacy notesRequired
Privacy notes are present.
Verify package integrity metadata
No package verification/checksum metadata.
Adopt in controlled steps based on the selected plan.
Run in isolated sandbox firstRequired
Use a constrained sandbox and observe behavior across multiple tasks.
Roll out graduallyRequired
Roll out to a small cohort before wider usage.
Set monitoring and fallback
Define rollback path and monitor errors after adoption.
Evidence readiness
Required evidence gates are covered (5/6 signals complete).
Source repository/provenance is listed.
Required in this preset
Review metadata is present.
Required in this preset
Safety notes are present.
Required in this preset
Privacy notes are present.
Optional in this preset
Package integrity metadata is missing.
Optional in this preset
Install payload is available.
Required in this preset
Required evidence gates are covered for this preset.
Decision timeline
5/6 steps complete with no blocking gaps for this preset.
triage
Source/provenance metadata is available.
triage
Review metadata is available.
verify
Safety notes are available.
verify
Privacy notes are available.
verify
Package integrity metadata is missing.
rollout
Install payload is available.
No required blockers for this timeline preset.
Safety & privacy surface
3 safety and 2 privacy notes across 4 risk areas. Review closely: credentials & tokens, permissions & scopes.
You are an advisor on using Anthropic Claude models inside GitHub Copilot and on moving work between GitHub Copilot and Claude Code. Ground every claim in GitHub Copilot's published docs and never invent model names, dates, prices, or APIs.You are an advisor agent for developers who use both GitHub Copilot and Claude Code. You help with two grounded things: (1) selecting an Anthropic Claude model inside GitHub Copilot Chat, and (2) deciding when to keep a task in Copilot versus moving it to Claude Code.
This is a user-created custom subagent (a system prompt you save as a Markdown file), not a built-in Claude Code command and not an official GitHub or Anthropic product. There is no "interoperability bridge", no automatic cross-platform router, and no shared-credential/auth-bridge tool. Copilot and Claude Code are separate products you switch between manually.
Save this as .claude/agents/copilot-claude-advisor.md in your project (or ~/.claude/agents/ for all projects):
---
name: copilot-claude-advisor
description: Advises on selecting Claude models in GitHub Copilot Chat and moving work between Copilot and Claude Code.
---
You are an advisor on using Anthropic Claude models inside GitHub Copilot and on
moving work between GitHub Copilot and Claude Code. Ground every claim in GitHub
Copilot's published docs. Never invent model names, release dates, prices, or APIs.
When the user names a task, recommend a Claude model available in Copilot and explain
when the same task is better done in Claude Code.
Then invoke it conversationally, e.g. "Which Claude model should I select in Copilot Chat for a multi-file refactor, and would Claude Code be a better fit?"
The dropdown shows the current model with a chevron; open it and pick a Claude model.
If a Claude model is missing from your picker, it is usually a plan or org-policy issue — confirm your Copilot subscription supports it and that an admin has enabled it.
This is editorial guidance, not a hard rule, and there is no tool that routes automatically:
Model names, plan tiers, and premium-request billing change. Before quoting a specific model or limit, open GitHub's supported models reference and the change the model for Copilot Chat guide listed in the sources. Do not state a model is available without confirming it on the live page.
Show that Claude Models in GitHub Copilot Advisor is listed on HeyClaude. Paste this Markdown into your README — it renders the badge and links back to this page.
[](https://heyclau.de/entry/agents/github-copilot-interop-bridge)Claude Models in GitHub Copilot Advisor side by side with 3 alternatives on trust, install, platform support, and disclosed safety notes — all from reviewed registry metadata.
1 trust signal differ across this comparison (Submitter).
| Field | A Claude Code subagent that helps you pick the right Anthropic Claude model in GitHub Copilot's model picker and decide when to switch to Claude Code instead. Open dossier | A Claude Code DevOps engineer persona grounded in GitHub Actions, Terraform, Docker, and Kubernetes. It writes and reviews CI/CD workflows, IaC, and deployment manifests using real tooling and documented commands. Open dossier | Source-backed agent for cleaning up open-source GitHub issue queues with reproducibility checks, labels, issue types, milestones, assignees, project views, duplicate links, and maintainer-safe response drafts. Open dossier | Source-backed agent for security review of open-source pull requests, including untrusted fork boundaries, GitHub Actions permissions, secret and code scanning, dependency review, provenance signals, and maintainer-owned merge recommendations. Open dossier |
|---|---|---|---|---|
| Next steps | ||||
| Trust | ||||
| Review status | ReviewedMaintainer reviewed | ReviewedMaintainer reviewed | ReviewedMaintainer reviewed | ReviewedMaintainer reviewed |
| Package trust | Package not verified | Package not verified | Package not verified | Package not verified |
| Source provenance | Source-backed | Source-backed | Source-backed | Source-backed |
| SubmitterDiffers | — | — | MkDev11 | MkDev11 |
| Install risk | Review first | Review first | Review first | Review first |
| Notes | Safety ✓ Privacy ✓ | Safety ✓ Privacy ✓ | Safety ✓ Privacy ✓ | Safety ✓ Privacy ✓ |
| Brand | — | |||
| Category | agents | agents | agents | agents |
| Source | Source-backed | Source-backed | Source-backed | Source-backed |
| Author | JSONbored | JSONbored | MkDev11 | MkDev11 |
| Added | 2025-10-25 | 2025-10-16 | 2026-06-05 | 2026-06-05 |
| Platforms | ||||
| Harness | ||||
| Source repo | — | — | — | — |
| Safety notes | ✓This is a prompt-only custom subagent; it executes no commands and installs nothing on its own. Model availability and premium-request billing depend on your GitHub Copilot plan and org/enterprise policy; selecting some Claude models may incur premium-request charges or require admin enablement. Copilot and Claude Code are separate products with separate authentication; the agent does not bridge or share credentials between them. | ✓This persona produces commands and pipelines that modify real infrastructure: terraform apply, kubectl apply/rollout, docker build/run, and GitHub Actions deploy jobs. Review terraform plan output and use a saved plan file (-out) before apply. Deployment workflows can mutate production. Use GitHub Environments with required reviewers and least-privilege GITHUB_TOKEN permissions; never auto-merge or auto-deploy without human review. kubectl rollout undo and terraform -destroy are destructive/irreversible against live state; confirm the target context/workspace before running. | ✓Treat label changes, assignments, milestone moves, project edits, issue closures, discussion conversions, and maintainer mentions as repository governance actions. Draft them first unless the maintainer explicitly authorizes mutation. Do not close issues only because they are old, vague, unpopular, or difficult to reproduce. Apply the repository's documented stale, support, duplicate, or not-planned policy and preserve appeal paths. Avoid making security, severity, ownership, roadmap, or legal claims from model inference alone. Escalate suspected vulnerabilities or abuse to the repository's private security and moderation process. When recommending automation, keep rate limits, notification volume, contributor trust, and maintainer review capacity in view. | ✓Treat public pull request code, generated artifacts, CI configuration, package scripts, and contributor-supplied test output as untrusted until a maintainer verifies the source diff and checks. Do not run untrusted fork code, package scripts, workflow changes, or reproduction commands with repository secrets, privileged tokens, or write permissions. Escalate before approval when the PR changes GitHub Actions permissions, pull_request_target behavior, release automation, dependency provenance, credential handling, auth, data deletion, or public security posture. Scanner output is evidence, not a final verdict. A clean scan does not replace diff review, owner signoff, exploitability reasoning, or current branch-protection checks. |
| Privacy notes | ✓Code and prompts sent through GitHub Copilot Chat are processed under GitHub Copilot's data-handling terms; code sent through Claude Code is processed under Anthropic's terms. These are distinct services with distinct policies. The agent itself stores no data; it only advises on model selection. | ✓Terraform state files (.tfstate) can contain secrets and resource identifiers; store them in a locked remote backend and never commit them to the repository. Do not place cloud credentials, kubeconfig contents, or registry tokens in workflow YAML or Dockerfiles; use GitHub Actions secrets/Environments and runtime secret stores instead. | ✓Issue queues can contain private logs, stack traces, screenshots, crash dumps, tokens, email addresses, customer names, hostnames, reproduction links, and unpublished roadmap details. Draft public comments with minimal necessary detail. Ask reporters to move secrets, credentials, vulnerability reports, or private customer data to an approved private channel instead of quoting it back into the issue. Saved replies, duplicate summaries, and project fields can reveal internal routing rules or maintainer availability. Keep internal notes separate from public-facing triage comments. Search queries, exported issue lists, and triage reports should be treated as contributor data and retained only as long as the repository policy allows. | ✓Security review notes can expose exploit details, secret values, private maintainer signals, abuse patterns, hidden CI logs, vulnerability reports, and embargoed project context. Redact secrets, tokens, private log lines, contributor abuse indicators, internal maintainer notes, and exploit steps before posting public PR comments. Keep public feedback actionable but minimal when a finding involves an unpatched vulnerability, suspected malicious contribution, private advisory, or credential exposure. |
| Prerequisites | — none listed | — none listed |
|
|
| Install | — | — | — | — |
| Config | — | — | — | — |
| Citations | ||||
| Claim | Unclaimed | Unclaimed | Unclaimed | Unclaimed |
DevOps subagent persona for GitHub Actions CI/CD, Terraform IaC, Docker, and Kubernetes. Real workflows and manifests, no predictive or self-healing claims.
Source-backed guides for putting this to work.
Loading live community signals…
A short, calm digest of reviewed Claude resources. Unsubscribe any time.