Cyrus Setup Claude AuthBLOCK
The Claude Code background agent for Linear, Slack, Github, GitLab etc. you deploy anywhere. Supports Codex, Cursor, Gemini, and Opencode harnesses too.
Overview
The Claude Code background agent for Linear, Slack, Github, GitLab etc. you deploy anywhere. Supports Codex, Cursor, Gemini, and Opencode harnesses too.
a792ab0e17bbOBSERVED · 2026-10-13Host compatibility
What the documentation claims. We have not run a compatibility test.
| Host | Status | Notes |
|---|---|---|
| claude-code | mentioned |
What it tells the agent
The instruction file, verbatim from the audited commit — this is the text the model reads, and the surface the audit's instruction layer examines. Quoted here so you can judge it without cloning anything.
--- name: cyrus-setup-claude-auth description: Configure Claude Code authentication for Cyrus — API key, OAuth token, or third-party provider. --- **CRITICAL: Never use `Read`, `Edit`, or `Write` tools on `~/.cyrus/.env` or any file inside `~/.cyrus/`. Use only `Bash` commands (`grep`, `printf >>`, etc.) to interact with env files — secrets must never be read into the conversation context.** # Setup Claude Auth Configures Claude Code credentials so Cyrus can run AI sessions. ## Step 1: Check Existing Auth Check if credentials are already configured: ```bash grep -c -E '^(ANTHROPIC_API_KEY|CLAUDE_CODE_OAUTH_TOKEN)=' ~/.cyrus/.env 2>/dev/null || echo "0" ``` If the count is >= 1, inform the user: > Claude Code authentication is already configured. Skipping this step. > To reconfigure, remove the existing key from `~/.cyrus/.env` and re-run this skill. Skip to completion. ## Step 2: Choose Auth Method Ask the user: > **How would you like to authenticate Claude Code?** > > 1. **Current account** (easiest) — use the credentials from your active `claude` CLI session > 2. **API Key** — from [console.anthropic.com](https://console.anthropic.com/) > 3. **Separate OAuth token** — run `claude setup-token` to generate a token for a specific account > 4. **Third-Party Provider** — Vertex AI, AWS Bedrock, Azure, etc. ## Step 3: Configure Credentials **CRITICAL: Secrets must NEVER appear in the conversation.** Do not explore `~/.claude/` looking for credential files. Use only the methods below. Detect the OS for the right clipboard command: ```bash uname -s ``` ### Option 1: Current Account Check if `claude` is authenticated on this machine: ```bash claude auth status ``` If authenticated, instruct the user to run `claude setup-token` which will output a token. To capture it safely, **append a placeholder line** to the env file, then open it in a GUI window so the user can paste the token: ```bash grep -q '^CLAUDE_CODE_OAUTH_TOKEN=' ~/.cyrus/.env || echo 'CLAUDE_CODE_OAUTH_TOKEN=' >> ~/.cyrus/.env ``` Then open the file in a separate window — use whatever GUI editor is available: ```bash # macOS: VS Code if available, otherwise TextEdit code --new-window ~/.cyrus/.env 2>/dev/null || open -a TextEdit ~/.cyrus/.env # Linux code ~/.cyrus/.env 2>/dev/null || xdg-open ~/.cyrus/.env ``` Tell the user: > 1. Run `claude setup-token` in a separate terminal > 2. Copy the token it outputs > 3. I've opened `~/.cyrus/.env` — find the `CLAUDE_CODE_OAUTH_TOKEN=` line > 4. Paste the token right after the `=` (no spaces, no newline) > 5. Save and close the file ### Option 2: API Key Instruct the user to copy their API key from the [Anthropic Console](https://console.anthropic.com/), then provide the appropriate command: **macOS:** ```bash printf 'ANTHROPIC_API_KEY=%s\n' "$(pbpaste)" >> ~/.cyrus/.env ``` **Linux:** ```bash printf 'ANTHROPIC_API_KEY=%s\n' "$(xclip -selection clipboard -o)" >> ~/.cyrus/.env ``` **Universal fallback:** ```bash read -s -p "Paste your Anthropic API key: " val && printf 'ANTHROPIC_API_KEY=%s\n' "$val" >> ~/.cyrus/.env && echo " ✓ Saved" ``` ### Option 3: Separate OAuth Token This is for when the user wants to generate a token for a different account than the one currently logged in (e.g., running `claude setup-token` on another machine). Append the placeholder, then open the file for the user to paste: ```bash grep -q '^CLAUDE_CODE_OAUTH_TOKEN=' ~/.cyrus/.env || echo 'CLAUDE_CODE_OAUTH_TOKEN=' >> ~/.cyrus/.env ``` ```bash # macOS code --new-window ~/.cyrus/.env 2>/dev/null || open -a TextEdit ~/.cyrus/.env # Linux code ~/.cyrus/.env 2>/dev/null || xdg-open ~/.cyrus/.env ``` Tell the user: > 1. On the other machine, run `claude setup-token` > 2. Copy the token it outputs > 3. I've opened `~/.cyrus/.env` — find `CLAUDE_CODE_OAUTH_TOKEN=` and paste the token after the `=` > 4. Save and close ### Option 4: Third-Party Provider Inform the user: > For third-party providers, you'll need to set provider-specific environment variables. > See [Third-Party Integrations](https://docs.anthropic.com/en/docs/claude-code/bedrock-vertex) for details. > > Common configurations: > > **AWS Bedrock:** > ``` > CLAUDE_CODE_USE_BEDROCK=1 > AWS_REGION=us-east-1 > ``` > > **Google Vertex AI:** > ``` > CLAUDE_CODE_USE_VERTEX=1 > CLOUD_ML_REGION=us-east5 > ANTHROPIC_VERTEX_PROJECT_ID=your-project-id > ``` Guide the user to add the appropriate variables to `~/.cyrus/.env` using the clipboard-to-env pattern. ## Step 4: Verify After the user runs the command, verify the key was written: ```bash grep -c -E '^(ANTHROPIC_API_KEY|CLAUDE_CODE_OAUTH_TOKEN|CLAUDE_CODE_USE_BEDROCK|CLAUDE_CODE_USE_VERTEX)=' ~/.cyrus/.env ``` If the count is 0, the credential was not saved. Ask the user to try again. ## Completion > ✓ Claude Code authentication configured.
Trust audit
BLOCKgrade D · trust 69/100 Do not install this without reading the findings. The audit found something that could harm you or your machine.
| Layer | What it checks | Result |
|---|---|---|
| L0 | Provenance & inventory | WARN |
| L1 | Static analysis of the code | NA |
| L2 | Instruction surface (what it tells the agent) | FAIL |
| L3 | Class-specific surface | PASS |
| L4 | Behavioural (sandbox) | SKIPPED |
What the source does
- Filesystem
- none-observed
- Network
- none-observed
- Shell
- none-observed
- Dependencies
- pinned
- Secrets in source
- none-found
Findings (6)
read -s -p "Paste your Anthropic API key: " val && printf 'ANTHROPIC_API_KEY=%s\n' "$val" >> ~/.cyrus/.env && echo " ✓ Saved"
.claude/skills/cyrus-setup
.claude/skills/cyrus-setup-claude-auth
.claude/skills/cyrus-setup-endpoint
.claude/skills/cyrus-setup-github
.claude/skills/cyrus-setup-gitlab
Gates applied: no_behavioural_pass.
a792ab0e17bbfull audit observations/trust-audit/skill/cyrusagents__cyrus-setup-claude-auth.json · Report an issue / request a re-scanAudit history
Every audit this skill has had.
| Date | Source | Verdict | Grade | Score | Change |
|---|---|---|---|---|---|
| 2026-10-13 | a792ab0e17bb | BLOCK | D | 69 | first audit |
Questions
What does the Cyrus Setup Claude Auth skill do?
The Claude Code background agent for Linear, Slack, Github, GitLab etc. you deploy anywhere. Supports Codex, Cursor, Gemini, and Opencode harnesses too.
Is Cyrus Setup Claude Auth safe to install?
No — not without reading the findings first. The audit graded it D (69/100) and found 1 critical or high issue in the source. Each one is listed on this page with the file and line it is on.
What can Cyrus Setup Claude Auth access on my machine?
The audit observed no filesystem, network or shell use at all in its source.
Which assistants does Cyrus Setup Claude Auth work with?
Its documentation mentions claude-code. That is what the text claims, not a compatibility test we ran.
How current is this page?
The grade is for one exact copy of the source (a792ab0e17bb), read on 2026-10-13. The repository is watched, and a new audit runs when it changes — this is the first audit.