Atlas / MCP servers / edgar-durand / codeam-cli

codeam-cliBLOCK

mcp/edgar-durand/codeagent-mobile-clients

Public source for CodeAgent Mobile clients: codeam-cli (npm), VS Code / Cursor / Windsurf extension, and JetBrains plugin. Control AI coding agents (Claude Code, Copilot, Cursor) from your phone.

Verdict
BLOCK
Grade
F
Trust score
41 /100
Exposed tools
20 11r · 8w · 1d
Transport
stdio · streamable-http
License
MIT
Stars
9
01

Overview

From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.

Public source for the client-side bridges of CodeAgent Mobile — the workflow-continuity layer for AI coding agents.

  • apps/cli — codeam-cli, the terminal bridge that wraps Claude Code in a PTY and streams it to any device
  • apps/vsc-plugin — extension for VS Code, Cursor, and Windsurf
  • apps/jetbrains-plugin — plugin for the IntelliJ family (IntelliJ IDEA, WebStorm, PyCharm, Rider, GoLand, ...)
What is CodeAgent Mobile? The supervision layer for agents already running in your IDE. Now that AI agents run for hours, not seconds, you no longer need to stare at a terminal — CodeAgent streams the session to your phone or web dashboard so you can approve diffs, redirect prompts, and stay in the loop async. Works with Claude Code, Copilot, Cursor, Windsurf, JetBrains AI, OpenAI Codex, Vertex AI, and any agent in your editor. Install once in your IDE, pair with a QR, your workflow follows you across devices. See codeagent-mobile.com.

Install (users)

CLI — codeam-cli

npm install -g codeam-cli
codeam pair   # generates a 6-character code
codeam        # starts Claude Code with mobile control

[](https://www.npmjs.com/package/codeam-cli)

VS Code / Cursor / Windsurf

JetBrains

Environment variables

Every client (CLI + both plugins) reads the same two env vars to pic

Read from source at commit c7bddfbf85e1OBSERVED · 2026-10-12
02

Connect

Built from this server's own package name, version and transport as found in its source — not copied from anyone's documentation, so it cannot drift against a page we do not control. Replace the environment placeholders with a token scoped to the least it needs.

claude-code
claude mcp add shared --env ANTHROPIC_API_KEY=${ANTHROPIC_API_KEY} --env ANTHROPIC_AUTH_TOKEN=${ANTHROPIC_AUTH_TOKEN} --env CLAUDE_CODE_OAUTH_TOKEN=${CLAUDE_CODE_OAUTH_TOKEN} --env CODEAM_AUTO_TOKEN=${CODEAM_AUTO_TOKEN} -- npx -y @codeam/[email protected]
claude-desktop
{
  "mcpServers": {
    "shared": {
      "command": "npx",
      "args": [
        "-y",
        "@codeam/[email protected]"
      ],
      "env": {
        "ANTHROPIC_API_KEY": "${ANTHROPIC_API_KEY}",
        "ANTHROPIC_AUTH_TOKEN": "${ANTHROPIC_AUTH_TOKEN}",
        "CLAUDE_CODE_OAUTH_TOKEN": "${CLAUDE_CODE_OAUTH_TOKEN}",
        "CODEAM_AUTO_TOKEN": "${CODEAM_AUTO_TOKEN}"
      }
    }
  }
}
03

Exposed tools (20)

11 read · 8 write · 1 destructive. Blast radius: 1 tool can delete or overwrite — an agent that can be talked into calling a tool can be talked into calling this one.

ToolRiskDescription
CoderreadImplements the task with tests, TDD-first.
DefaultreadNormal editing
QAreadVerifies every acceptance criterion end to end and writes the final report.
ReviewerreadSkeptical review in a fresh context — finds and fixes what the coder missed.
SpecifierreadTurns the task into testable acceptance criteria before any code.
call_toolwriteInvoke one integration tool by its
clear_highlightsdestructiveRemove every mark you drew in the Preview with highlight_element.
datareadRead documents from a table (most-recent first). Provide the table name.
deploywriteDeploy this repo\
describe_toolreadReturn the full JSON input schema of one tool, given its
highlight_elementwritePoint at an element in the running Preview: the user sees your cursor move to it and a labelled box drawn
list_integrationsreadList the integrations attached to this session (Jira, Linear, ClickUp, Postman, ...) with how many tools each exposes and whether it is available. Call this first when the user asks what you can do with an external service.
preview_statusreadCurrent state of the Preview: running (with its URL), starting, error (with the reason) or idle.
run_mutationwriteRun a Convex mutation function by its path with JSON args (writes data).
run_querywriteRun a read-only Convex query function by its path (e.g.
schemareadGet the deployed schema (table definitions + validators) of the Convex deployment.
search_toolsreadSearch the tools of the attached integrations by keywords (e.g.
start_previewwriteOpen this project
stop_previewwriteStop the Preview (dev server + public URL). Only when the user asks. Never to apply code changes — the dev
tableswriteList the tables in the connected Convex deployment (name + id). Start here to discover the schema.
04

Trust audit

BLOCKgrade F · trust 41/100 Do not install this without reading the findings. The audit found something that could harm you or your machine.

LayerWhat it checksResult
L0Provenance & inventoryFAIL
L1Static analysis of the codeFAIL
L2Instruction surface (what it tells the agent)PASS
L3Class-specific surfaceWARN
L4Behavioural (sandbox)SKIPPED

What the source does

Filesystem
declared (4 observation(s))
Network
declared (2 observation(s))
Shell
declared (7 observation(s))
Dependencies
not all pinned
Secrets in source
found

Findings (25)

HIGHInventory / provenance · inv.suspicious_name · CWE-1104
apps/cli/src/lib/payload.ts
payload.ts
Why it matters. member named after an attack tool
Fix. remove or justify
HIGHInventory / provenance · inv.suspicious_name · CWE-1104
apps/cli/src/skills/persist-from-payload.ts
persist-from-payload.ts
Why it matters. member named after an attack tool
Fix. remove or justify
HIGHCode injection · code.eval_exec · CWE-78, CWE-94, CWE-95
apps/cli/src/scratch/push-clean.ts:43
const { stdout: keys } = await exec(
Why it matters. evaluates text as code
Fix. remove; use a parser or a dispatch table
HIGHCode injection · code.eval_exec · CWE-78, CWE-94, CWE-95
apps/cli/src/scratch/push-clean.ts:62
await exec(
Why it matters. evaluates text as code
Fix. remove; use a parser or a dispatch table
HIGHCode injection · code.eval_exec · CWE-78, CWE-94, CWE-95
apps/cli/src/services/providers/types.ts:149
exec(workspaceId: string, command: string): Promise<ExecResult>;
Why it matters. evaluates text as code
Fix. remove; use a parser or a dispatch table
HIGHCode injection · code.eval_exec · CWE-78, CWE-94, CWE-95
apps/vsc-plugin/src/services/file-watcher.service.ts:558
exec(
Why it matters. evaluates text as code
Fix. remove; use a parser or a dispatch table
HIGHNetwork egress · net.env_exfil · CWE-200, CWE-319
apps/cli/src/agents/coderabbit/installer.ts:306
{ ...process.env } ... curl
Why it matters. reads secrets in the same file that sends data out
HIGHHard-coded secrets · secret.db_uri · CWE-798, CWE-321
apps/cli/src/services/preview/provision-deps.ts:45
'DATABASE_URL=postgresql://postgres:postgres@localhost:5432/app',
MEDIUMInventory / provenance · inv.binary · CWE-1104
apps/jetbrains-plugin/gradle/wrapper/gradle-wrapper.jar
gradle-wrapper.jar
Why it matters. a compiled or binary member cannot be reviewed from source
Fix. ship source, or explain the binary in the README
MEDIUMNetwork egress · net.beacon_words · CWE-200, CWE-319
packages/shared/src/skills/agent-standard.ts:65
- **Never expose or exfiltrate** secrets, credentials, tokens, or customer data, and never print a credential's value.
MEDIUMHard-coded secrets · secret.anthropic · CWE-798, CWE-321
apps/cli/__tests__/agents/claude.setupToken.test.ts:9
'sk-ant-oat01-AbC123_def-456GHI_jkl789-MNOpqr',
MEDIUMHard-coded secrets · secret.anthropic · CWE-798, CWE-321
apps/cli/__tests__/agents/claude.setupToken.test.ts:12
expect(extractSetupTokenFromOutput(out)).toBe('sk-ant-oat01-AbC123_def-456GHI_jkl789-MNOpqr');
MEDIUMHard-coded secrets · secret.anthropic · CWE-798, CWE-321
apps/cli/__tests__/lib/provider-routing.test.ts:189
'sk-ant-api03-ABCDEFGHIJKLMNOP Bearer abcdefghijklmnopqrstuvwxyz ' +
MEDIUMHard-coded secrets · secret.db_uri · CWE-798, CWE-321
apps/cli/__tests__/preview/dotenv.test.ts:22
{ key: 'URL', value: 'postgres://u:p@h:5432/db?x=1' },
MEDIUMHard-coded secrets · secret.db_uri · CWE-798, CWE-321
apps/cli/__tests__/preview/dotenv.test.ts:54
{ key: 'URL', value: 'postgres://u:p@h:5432/db?x=1' },
MEDIUMHard-coded secrets · secret.generic · CWE-798, CWE-321
apps/cli/__tests__/host-agent.test.ts:3658
const secret = 'super-secret-enroll-token';
MEDIUMHard-coded secrets · secret.generic · CWE-798, CWE-321
apps/cli/__tests__/scratch/save-github.real-git.test.ts:20
const TOKEN = 'gho_realgit_secret_123';
MEDIUMFilesystem / path · mcp.destructive_tools · CWE-22, CWE-59
clear_highlights
Why it matters. 1 tool(s) can delete or overwrite
Fix. prefer a read-only mode or scoped tokens; the page states the blast radius
LOWInventory / provenance · inv.binary · CWE-1104
apps/cli/__tests__/fixtures/codex-multiline-reply.bin
codex-multiline-reply.bin
Why it matters. a compiled or binary member cannot be reviewed from source
Fix. ship source, or explain the binary in the README
LOWInventory / provenance · inv.hidden_file · CWE-1104
.gitmessage
.gitmessage
Why it matters. hidden member outside the usual dotfiles
Fix. review its purpose
LOWInventory / provenance · inv.hidden_file · CWE-1104
apps/vsc-plugin/.vscodeignore
.vscodeignore
Why it matters. hidden member outside the usual dotfiles
Fix. review its purpose
LOWInventory / provenance · inv.suspicious_name · CWE-1104
apps/cli/__tests__/preview/origin-payload.test.ts
origin-payload.test.ts
Why it matters. member named after an attack tool
Fix. remove or justify
LOWInventory / provenance · inv.suspicious_name · CWE-1104
apps/cli/__tests__/services/payload-parsing.test.ts
payload-parsing.test.ts
Why it matters. member named after an attack tool
Fix. remove or justify
LOWInsecure crypto · crypto.weak_hash · CWE-327, CWE-338
apps/cli/__tests__/agents/cursor.bridge.test.ts:26
return createHash('md5').update(s).digest('hex');
LOWInsecure crypto · crypto.weak_hash · CWE-327, CWE-338
apps/cli/src/agents/cursor/history.ts:68
const computed = path.join(chatsRoot, createHash('md5').update(cwd).digest('hex'), sessionId);

Gates applied: no_behavioural_pass.

Audited 2026-10-12 · audit v0.4.1 · source sha c7bddfbf85e1full audit observations/trust-audit/mcp-server/edgar-durand__codeagent-mobile-clients.json · Report an issue / request a re-scan
05

Audit history

Every audit this server has had. A grade with a past is a grade somebody is still checking.

DateSourceVerdictGradeScoreChange
2026-10-12c7bddfbf85e1BLOCKF41first audit
06

Questions

What is the codeam-cli MCP server?

Public source for CodeAgent Mobile clients: codeam-cli (npm), VS Code / Cursor / Windsurf extension, and JetBrains plugin. Control AI coding agents (Claude Code, Copilot, Cursor) from your phone.

What tools does codeam-cli expose?

20 in total: 11 read-only, 8 that write, and 1 that can delete or overwrite (clear_highlights). Every one is listed on this page with its risk.

Is codeam-cli safe to connect to an agent?

No — not without reading the findings first. The audit graded it F (41/100) and found 8 critical or high issues in the source. Each one is listed on this page with the file and line it is on. Separately from the audit: 1 of its tools can destroy data, so scope the token you give it to what you actually need.

What credentials does codeam-cli need?

It reads ANTHROPIC_API_KEY, ANTHROPIC_AUTH_TOKEN, CLAUDE_CODE_OAUTH_TOKEN, CODEAM_AUTO_TOKEN, CODEAM_ENROLL_TOKEN, CODEAM_MCP_PLUGIN_TOKEN, CODEAM_MCP_POLL_SECRET, CODEAM_VERCEL_BYPASS, DEEPSEEK_API_KEY, FAKE_TOKEN, GEMINI_API_KEY and GROQ_API_KEY from the environment. Give it a token scoped to the least it needs — an agent that can be talked into calling a tool can be talked into calling it with your credentials.

How does codeam-cli run?

It speaks stdio and streamable-http, so it runs as a local process your client starts. It is published on npm as @codeam/shared at 2.2.0.

How current is this page?

The grade is for one exact copy of the source (c7bddfbf85e1), read on 2026-10-12. The repository is watched and re-audited when it changes.

Advertisement