Fix Git error: fatal: Could not read from remote repository

What this error means

Git tried to contact the remote (origin) but could not authenticate, reach it, or find the repository. The most common causes are:

• Wrong remote URL (typo, protocol mismatch)
• Missing or unusable SSH key
• Insufficient permissions on the remote
• HTTPS credentials rejected (need a token)
• Network, proxy, or host key verification issues

Quickstart (fast checklist)

• Verify the remote: git remote -v
• Test connectivity:
  • SSH: ssh -T [email protected] (or your host)
  • HTTPS: git ls-remote https://github.com/example/acme.git
• Fix the URL: git remote set-url origin <correct-url>
• SSH: load key and try again: ssh-add -l || ssh-add ~/.ssh/id_ed25519
• HTTPS: use a Personal Access Token instead of a password
• If CI/CD: ensure the key/secret and known_hosts are present

Minimal working example (local bare remote)

Use a local bare repo as a remote to confirm Git works end-to-end on your machine.

set -euo pipefail
# 1) Create a bare "remote"
rm -rf /tmp/git-remote-demo
mkdir -p /tmp/git-remote-demo
git init --bare /tmp/git-remote-demo/remote.git

# 2) Create a repo and push to the local remote
rm -rf /tmp/git-wrk && mkdir -p /tmp/git-wrk && cd /tmp/git-wrk
git init
printf "hello\n" > README.md
git add README.md
git commit -m "init"
git checkout -b main
git remote add origin /tmp/git-remote-demo/remote.git
git push -u origin main

# Success here proves your Git install and basic permissions are fine.

If this works but your real remote fails, the issue is with remote URL, auth, or network.

Step-by-step fix

1. Confirm the remote URL

git remote -v
# If wrong, set it:
# SSH example
git remote set-url origin [email protected]:example/acme.git
# HTTPS example
git remote set-url origin https://github.com/example/acme.git

1. Check repository existence and your access

• Ensure the project exists on the server and your account or deploy key has read/write rights.
• If you see “Repository not found”, verify org/team membership and repo visibility.

1. SSH authentication (common case)

# List loaded keys
ssh-add -l || eval "$(ssh-agent -s)" && ssh-add ~/.ssh/id_ed25519

# Test host auth (replace host as needed)
ssh -T [email protected]

# If the wrong key is chosen, create an SSH config:

# ~/.ssh/config
Host github.com
  HostName github.com
  User git
  IdentityFile ~/.ssh/id_ed25519
  IdentitiesOnly yes

# Fix key permissions if needed
chmod 700 ~/.ssh
chmod 600 ~/.ssh/id_ed25519 ~/.ssh/id_rsa 2>/dev/null || true

# If host key changed or missing
ssh-keygen -R github.com
ssh -o StrictHostKeyChecking=accept-new [email protected] || true

# Increase SSH verbosity when testing Git
GIT_SSH_COMMAND='ssh -v' git ls-remote origin

1. HTTPS authentication

• Many providers require a Personal Access Token (PAT) instead of a password.
• When prompted for password, paste the PAT.

# Update to HTTPS if preferred
git remote set-url origin https://github.com/example/acme.git

# Clear bad cached creds and retry to prompt for new creds
git credential reject <<<'protocol=https\nhost=github.com\n'
# Optional: cache credentials for this session
git config --global credential.helper cache

1. Proxies, firewalls, and corporate networks

# Inspect proxy settings
env | grep -i proxy || true
git config --global --get http.proxy || true

# Temporarily disable for test
unset http_proxy https_proxy no_proxy
GIT_SSH_COMMAND='ssh -o ConnectTimeout=10' git ls-remote origin

• If SSH is blocked outbound, switch to HTTPS.
• If HTTPS is intercepted, install the corporate CA certificate and configure Git/cURL to trust it.

1. CI/CD runners (DevOps)

• Provide the private key via secret variables; add known_hosts for the Git host.
• Use a deploy key or a bot user with the least privileges required.
• Non-interactive environments need ssh-agent and key loading commands in the job.

1. Submodules

• Each submodule has its own remote. Run git submodule foreach 'git remote -v' and fix URLs and auth per submodule.

1. Retry the operation

git fetch --prune
git push -u origin HEAD

Error-to-cause quick map

Pitfalls to avoid

• Mixing multiple SSH keys without an SSH config; the wrong key gets used.
• Using deprecated keys (DSA, weak RSA) blocked by providers; prefer ed25519.
• Wrong protocol for policy (company enforces HTTPS-only or SSH-only).
• CI jobs missing known_hosts, causing strict host checks to fail.
• Windows: ensure you use the same SSH (Git Bash vs Windows OpenSSH) consistently, or specify core.sshCommand.

Performance notes

• Prefer SSH multiplexing to speed repeated operations:

# ~/.ssh/config
Host github.com
  HostName github.com
  User git
  IdentityFile ~/.ssh/id_ed25519
  ControlMaster auto
  ControlPersist 5m
  ControlPath ~/.ssh/cm-%r@%h:%p

• Use shallow and partial clones for large repos:

git clone --depth 1 --filter=blob:none [email protected]:example/acme.git

• Parallelize fetches:

git -c fetch.parallel=4 fetch --all --prune

• Reduce SSH/TLS handshake cost in CI by reusing the workspace or enabling caching between steps.

FAQ

• Why does SSH fail but HTTPS works?

  • Likely port 22 blocked or SSH key not available. Use HTTPS or provide the key.

• Do I need a PAT for HTTPS?

  • Yes on most providers if passwords are disabled. Generate a token with the needed scopes.

• It works locally but fails in CI—why?

  • Missing key/known_hosts or different URL. Add ssh-agent steps and known_hosts to the pipeline.

• Can I safely delete known_hosts?

  • Do not delete the whole file. Remove only the stale host entry with ssh-keygen -R host and re-add.