CVE-2026-27626

Summary

OliveTin's shell mode safety check (checkShellArgumentSafety) blocks several dangerous argument types but not password. A user supplying a password-typed argument can inject shell metacharacters that execute arbitrary OS commands. A second independent vector allows unauthenticated RCE via webhook-extracted JSON values that skip type safety checks entirely before reaching sh -c.

Details

Vector 1 — password type bypasses shell safety check (PR:L)

service/internal/executor/arguments.go has two gaps:

// Line 198-199 — TypeSafetyCheck returns nil (no error) for password type
case "password":
    return nil  // accepts ANY string including ; | ` $()
// Line 313 — checkShellArgumentSafety blocks dangerous types but not password
unsafe := map[string]bool{
    "url":                      true,
    "email":                    true,
    "raw_string_multiline":     true,
    "very_dangerous_raw_string": true,
    // "password" is absent — not blocked
}

Shell execution at service/internal/executor/executor_unix.go:18:

exec.CommandContext(ctx, "sh", "-c", finalParsedCommand)

A user supplies a password argument value of '; id; echo ' → sh -c interprets the shell metacharacters → arbitrary command execution.

This is not the "admin already has access" pattern: OliveTin explicitly enforces an admin/user boundary where admins define commands and users only supply argument values. The password type is the documented, intended mechanism for user-supplied sensitive values. The safety check exists precisely to prevent users from escaping this boundary — password is the one type it fails to block.

Vector 2 — Webhook JSON extraction skips TypeSafetyCheck entirely (PR:N)

service/internal/executor/handler.go:153-157 extracts arbitrary key-value pairs from webhook JSON payloads and injects them into ExecutionRequest.Arguments. These webhook-extracted arguments have no corresponding config-defined ActionArgument entry, so parseActionArguments() in arguments.go finds no type to check against and skips TypeSafetyCheck entirely. The values are templated directly into the shell command and passed to sh -c.

Example: an admin command template git pull && echo {{ git_message }} with Shell mode enabled. A webhook POST with {"git_message": "x; id"} injects id into the shell command. The webhook endpoint is unauthenticated by default (authType: none in default config).

PoC

## Vector 1 — authenticated user with password-type argument
curl -X POST http://localhost:1337/api/StartAction \
  -H "Content-Type: application/json" \
  -d '{"actionId": "run-command", "arguments": [{"name": "pass", "value": "'; id; echo '"}]}'
## Vector 2 — unauthenticated webhook
curl -X POST http://localhost:1337/webhook/git-deploy \
  -H "Content-Type: application/json" \
  -d '{"git_message": "x; id #", "git_author": "attacker"}'

Confirmed on jamesread/olivetin:latest (3000.10.0), 3/3 runs. Both vectors produced uid=1000(olivetin) output and arbitrary file write to /tmp/pwned.

Impact

• Vector 1: Any authenticated user (registration enabled by default, authType: none by default) can execute arbitrary OS commands on the OliveTin host with the permissions of the OliveTin process.
• Vector 2: Unauthenticated attacker can achieve the same if the instance receives webhooks from external sources, which is a primary OliveTin use case.

Combined: unauthenticated RCE on any OliveTin instance using Shell mode with webhook-triggered actions.