Back to skill

Security audit

Free Ride - Unlimited free AI

Security checks for vulnerabilities and agentic risk

Overview

The skill is rated suspicious because scanners report broad configuration changes and background persistence concerns; the review is incomplete because some referenced paths were not resolved.

Consider the suspicious rating carefully before installing. Review how model routing and background watching work, what local configuration and credentials the skill reads, and which external providers receive requests. The scan is partial because several referenced paths could not be resolved, and the requests dependency version is unknown.

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Memory PoisoningPersistent Context Injection, Context Window Stuffing, Memory Manipulation
  • Rogue AgentSelf-Modification, Session Persistence
Findings (27)

Credential Access

High
Category
Privilege Escalation
Confidence
60% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · README.md (reported line 22)May include surrounding context.

md
>
> ```bash
> uv tool install freeride-gateway==0.4.0a28   # or: pipx install freeride-gateway==0.4.0a28
> freeride init                                  # asks for keys, stores them in ~/.freeride/.env
> freeride serve
> freeride bind openclaw                         # points OpenClaw at the gateway
> ```

Credential Access

High
Category
Privilege Escalation
Confidence
60% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · SKILL.md (reported line 33)May include surrounding context.

md
- api.cerebras.ai (only with --providers cerebras)
reads:
  - ~/.openclaw/openclaw.json
  - ~/.openclaw/.env (API keys, read only)
writes:
  - ~/.openclaw/openclaw.json (keys: agents.defaults.model.primary, agents.defaults.model.fallbacks, agents.defaults.models.<model-ref>; and auth.profiles.openrouter:default only with --setup-auth after confirmation)
  - ~/.openclaw/openclaw.json.freeride-tmp (temporary, replaced atomically)

Credential Access

High
Category
Privilege Escalation
Confidence
60% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · SKILL.md (reported line 45)May include surrounding context.

md
- ~/.openclaw/ (created if missing)
never-writes:
  - API keys, anywhere
  - ~/.openclaw/.env
  - auth-profiles.json
runs:
  - openclaw plugins list --json (read only, only with --providers)

Credential Access

High
Category
Privilege Escalation
Confidence
60% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · main.py (reported line 39)May include surrounding context.

python
# Copy of openclaw.json taken before every FreeRide write.
BACKUP_FILE = Path.home() / ".openclaw" / "openclaw.json.freeride-bak"
# OpenClaw's recommended place for provider keys. FreeRide only reads it.
DOTENV_PATH = Path.home() / ".openclaw" / ".env"

# OpenRouter app-attribution headers — applied to every request so all
# FreeRide traffic shows up under one identity on OpenRouter's App Activity

Credential Access

High
Category
Privilege Escalation
Confidence
60% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · main.py (reported line 810)May include surrounding context.

python
"""List available free models ranked by quality."""
    if not get_api_keys():
        print("Error: OPENROUTER_API_KEY not set")
        print("Add OPENROUTER_API_KEY=<your key> to ~/.openclaw/.env in your editor,")
        print("or export it in your shell. Get a free key at: https://openrouter.ai/keys")
        sys.exit(1)

Credential Access

High
Category
Privilege Escalation
Confidence
60% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · providers.py (reported line 123)May include surrounding context.

python
# ============== Keys (read only) ==============

def read_dotenv(path: Path) -> dict:
    """Parse a simple KEY=VALUE .env file. Never written by FreeRide."""
    values = {}
    try:
        text = Path(path).read_text()

Credential Access

High
Category
Privilege Escalation
Confidence
60% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · skill.json (reported line 76)May include surrounding context.

json
"network": ["openrouter.ai", "api.groq.com", "generativelanguage.googleapis.com", "api.cerebras.ai"],
  "reads": [
    "~/.openclaw/openclaw.json",
    "~/.openclaw/.env (keys only, never written)"
  ],
  "writes": [
    "~/.openclaw/openclaw.json (only the configKeys below)",

Credential Access

High
Category
Privilege Escalation
Confidence
60% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · skill.json (reported line 90)May include surrounding context.

json
],
  "neverWrites": [
    "API keys, anywhere",
    "~/.openclaw/.env",
    "auth-profiles.json"
  ],
  "runs": [

Memory Manipulation

High
Category
Memory Poisoning
Confidence
90% confidence
Finding

Skill manipulates agent memory, state, or stored context. Memory corruption can alter personality, override safety rules, or cause unpredictable behavior.

Content

Scanner excerpt · watcher.py (reported line 46)May include surrounding context.

python
def _atomic_write(path: Path, content: str):
    """Write atomically via tmp + rename so a crash mid-write can't corrupt state."""
    path.parent.mkdir(parents=True, exist_ok=True)
    tmp = path.with_suffix(path.suffix + ".tmp")
    tmp.write_text(content)

Session Persistence

Medium
Category
Rogue Agent
Confidence
65% confidence
Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Content

Scanner excerpt · README.md (reported line 244)May include surrounding context.

md
freeride-watcher

# Background, persistent across logout
nohup freeride-watcher > ~/.openclaw/freeride-watcher.log 2>&1 &

# One-off check (no loop)
freeride-watcher --once

Session Persistence

Medium
Category
Rogue Agent
Confidence
65% confidence
Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Content

Scanner excerpt · SKILL.md (reported line 212)May include surrounding context.

md
freeride-watcher

# Persistent background (writes the log file ~/.openclaw/freeride-watcher.log)
nohup freeride-watcher > ~/.openclaw/freeride-watcher.log 2>&1 &

# One-shot check (no loop)
freeride-watcher --once

Unrestricted Tool Access

Medium
Category
Excessive Agency
Confidence
80% confidence
Finding

Skill grants unrestricted tool access without appropriate constraints. An agent with unfettered tool access can perform arbitrary actions including file modification, network requests, and code execution.

Content

Scanner excerpt · README.md (reported line 301)May include surrounding context.

md
**I'm upgrading from 1.0.x. What changes for me?**

- The first time you run any command, FreeRide works out who owns your primary. An OpenRouter `:free` model (or `openrouter/free`) counts as FreeRide's; anything else, including a paid model, counts as yours and is saved for `freeride restore`. If you hand-picked a free model and want it left alone, run `freeride auto -f`.
- The watcher and `freeride rotate` stop touching a primary you set.
- If an older FreeRide already replaced your paid primary, it was never saved, so `restore` can't bring it back. Set it again yourself (and then `freeride auto -f` to keep it).
- Nothing else changes unless you opt into extra providers.

Skill declares no tool scope ('permissions' or 'allowed-tools') but code capabilities were detected: env, file_read, file_write, network, shell.

Medium
Category
MCP Least Privilege
Confidence
70% confidence
Finding

Without declared permissions the skill's intent is opaque and cannot be validated.

Content

No source excerpt is available for this finding.

Activation includes generic mentions of rate limits or model switching without requiring an OpenClaw configuration request.

Medium
Category
Trigger Abuse
Confidence
90% confidence
Finding

The activation criteria include generic discussions of free AI, rate limits, and model switching rather than requiring a request to modify OpenClaw. Combined with the workflow that runs configuration changes and restarts the gateway when a user wants free AI, this can turn an informational request into unintended persistent routing changes and service interruption.

Content

No source excerpt is available for this finding.

Session Persistence

Medium
Category
Rogue Agent
Confidence
65% confidence
Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Content

Scanner excerpt · SKILL.md (reported line 41)May include surrounding context.

md
- ~/.openclaw/.freeride-state.json (+ .freeride-tmp sibling)
  - ~/.openclaw/.freeride-cache.json
  - ~/.openclaw/.freeride-watcher-state.json (+ .tmp sibling)
  - ~/.openclaw/freeride-watcher.log (only via the suggested nohup redirect)
  - ~/.openclaw/ (created if missing)
never-writes:
  - API keys, anywhere

Session Persistence

Medium
Category
Rogue Agent
Confidence
60% confidence
Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Content

Scanner excerpt · SKILL.md (reported line 165)May include surrounding context.

md
Everything else (gateway, channels, plugins, env, customInstructions, named agents) is preserved.

Before every write FreeRide copies the file to `~/.openclaw/openclaw.json.freeride-bak` and writes the new file atomically. If `openclaw.json` has JSON5 comments or trailing commas, or uses `$include`, FreeRide refuses to write it and says why.

Other files FreeRide writes, all under `~/.openclaw/`:

Ownership inference can rotate user-configured models despite the documented promise to leave them untouched.

Medium
Category
Excessive Agency
Confidence
98% confidence
Finding

When ownership state is missing or unreadable, _infer_state treats any existing OpenRouter free-model primary as FreeRide-owned solely from its identifier. Consequently, rotate can probe and replace a user-configured primary without explicit delegation, potentially changing which model receives subsequent agent traffic and failing to preserve the original primary for restoration.

Content

No source excerpt is available for this finding.

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
75% confidence
Finding

Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Content

Scanner excerpt · main.py (reported line 684)May include surrounding context.

python
def confirm_setup_auth(config: dict, assume_yes: bool) -> bool:
    """Explain exactly what --setup-auth writes and get consent first.
    Returns False if the user declined or no consent could be asked for."""
    profiles = (config.get("auth") or {}).get("profiles") or {}
    if AUTH_PROFILE_ID in profiles:
        print(f"Auth profile {AUTH_PROFILE_ID} already exists; --setup-auth will not change it.")

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
75% confidence
Finding

Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Content

Scanner excerpt · main.py (reported line 1421)May include surrounding context.

python
switch_parser.add_argument("--setup-auth", action="store_true",
                              help="Also add the auth.profiles[\"openrouter:default\"] entry (asks first)")
    switch_parser.add_argument("--yes", "-y", action="store_true",
                              help="Confirm --setup-auth without asking")
    _add_provider_args(switch_parser)

    # auto command

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
75% confidence
Finding

Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Content

Scanner excerpt · main.py (reported line 1433)May include surrounding context.

python
auto_parser.add_argument("--setup-auth", action="store_true",
                            help="Also add the auth.profiles[\"openrouter:default\"] entry (asks first)")
    auto_parser.add_argument("--yes", "-y", action="store_true",
                            help="Confirm --setup-auth without asking")
    _add_provider_args(auto_parser)

    # status command

External Transmission

Medium
Category
Data Exfiltration
Confidence
60% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · providers.py (reported line 66)May include surrounding context.

python
name="groq",
        prefix="groq",
        env_vars=("GROQ_API_KEY",),
        base_url="https://api.groq.com/openai/v1",
        plugin_package="@openclaw/groq-provider",
        signup_url="https://console.groq.com/keys",
    ),

External Transmission

Medium
Category
Data Exfiltration
Confidence
60% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · providers.py (reported line 85)May include surrounding context.

python
name="cerebras",
        prefix="cerebras",
        env_vars=("CEREBRAS_API_KEY",),
        base_url="https://api.cerebras.ai/v1",
        plugin_package="@openclaw/cerebras-provider",
        signup_url="https://cloud.cerebras.ai",
    ),

subprocess module call

Medium
Category
Dangerous Code Execution
Confidence
70% confidence
Finding

subprocess module calls execute external commands. Without careful input validation, this enables command injection.

Content

Scanner excerpt · providers.py (reported line 180)May include surrounding context.

python
if not binary:
        return None
    try:
        proc = subprocess.run([binary] + args, capture_output=True, text=True, timeout=60)
    except (OSError, subprocess.SubprocessError):
        return None
    if proc.returncode != 0:

Session Persistence

Medium
Category
Rogue Agent
Confidence
65% confidence
Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Content

Scanner excerpt · skill.json (reported line 85)May include surrounding context.

json
"~/.openclaw/.freeride-state.json (+ .freeride-tmp sibling)",
    "~/.openclaw/.freeride-cache.json",
    "~/.openclaw/.freeride-watcher-state.json (+ .tmp sibling)",
    "~/.openclaw/freeride-watcher.log (only via the suggested nohup redirect)",
    "~/.openclaw/ (created if missing)"
  ],
  "neverWrites": [

Scope Creep

Low
Category
Excessive Agency
Confidence
70% confidence
Finding

Skill's behavior or capabilities extend beyond its stated purpose. Scope creep allows an agent to perform actions unrelated to its documented functionality, increasing the attack surface.

Content

Scanner excerpt · LICENSE (reported line 12)May include surrounding context.

text
permit persons to whom the Software is furnished to do so.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED,
INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A
PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT
HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE