Back to skill

Security audit

CrowTerminal

Security checks for vulnerabilities and agentic risk

Overview

CrowTerminal is a disclosed third-party persistent memory API for creator analytics, but users should treat uploaded creator data and webhooks as sensitive.

Install/use this only if you are comfortable sending creator strategy, retention, watch-time, demographic, and webhook event data to CrowTerminal. Redact sensitive or personal data, use the sandbox first, pin SDK versions if installing them, and confirm retention/deletion and webhook security expectations before using real client data.

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • 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
Findings (1)

T08 · Insecure Dependencies

Warning
Location
SKILL.md:279
Finding
Unpinned Third-Party SDK Installation## Vulnerability Details **File Location**: `SKILL.md:279` **Vulnerability Type**: Insecure dependency installation **Risk Level**: Medium **Vulnerable Code Snippet**: ```markdown - **SDKs**: Python (`pip install crowterminal`), TypeScript (`npm install crowterminal`) ``` ### Technical Analysis The documentation recommends installing the `crowterminal` packages from public Python and npm registries without fixed versions, hashes, lockfiles, or integrity verification. Consequently, users following these commands will install whichever package release the registries resolve at installation time rather than a version reviewed with this Skill. Package installation can execute package-controlled build or lifecycle logic. The project file also does not provide evidence that the registry artifacts are reproducibly built from, or cryptographically tied to, the referenced source repository. This creates a supply-chain risk if a publisher account, package, registry, or future release is compromised. ### Attack Path 1. An attacker compromises the `crowterminal` registry package, its publisher account, or its release pipeline. 2. The attacker publishes a malicious version containing installation-time or runtime payloads. 3. A user follows `SKILL.md:279` and runs `pip install crowterminal` or `npm install crowterminal`. 4. The package manager resolves and downloads the attacker-controlled current release. 5. Malicious installation, build, lifecycle, import-time, or runtime code executes in the user's environment. ### Impact Assessment Successful exploitation could execute arbitrary code with the privileges of the user or automation account running the package manager. Depending on that environment, the payload could access application files, environment variables such as `CROWTERMINAL_API_KEY`, developer credentials, creator data, and network-accessible services. It could also modify user-writable files or dependencies. No malici ...[truncated 134 chars]
Remediation
## Remediation Suggestions 1. Replace floating installation commands with exact, reviewed versions, such as `crowterminal==X.Y.Z` for Python and `crowterminal@X.Y.Z` for npm. 2. Provide lockfiles and integrity metadata. For Python, publish a requirements file with cryptographic hashes and install with `pip --require-hashes`. For npm, commit and use a reviewed lockfile with `npm ci`. 3. Document the packages' verified official registry pages and their relationship to the referenced source repository. 4. Verify package signatures, provenance attestations, and checksums where supported. 5. Review package lifecycle and build scripts before adoption. Disable npm lifecycle scripts with `--ignore-scripts` when they are not required. 6. Install dependencies in a least-privileged, isolated virtual environment or container, and avoid exposing unrelated credentials during installation. 7. Establish an update-review process rather than allowing installations to resolve automatically to newly published releases.
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (16)

External Transmission

Medium
Category
Data Exfiltration
Content
clawdbot:
    category: productivity
    requires_api_key: true
    api_key_url: https://api.crowterminal.com/api/agent/register
requires:
  env:
    - CROWTERMINAL_API_KEY
Confidence
50% confidence
Finding
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

External Transmission

Medium
Category
Data Exfiltration
Content
clawdbot:
    category: productivity
    requires_api_key: true
    api_key_url: https://api.crowterminal.com/api/agent/register
requires:
  env:
    - CROWTERMINAL_API_KEY
Confidence
50% confidence
Finding
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

External Transmission

Medium
Category
Data Exfiltration
Content
clawdbot:
    category: productivity
    requires_api_key: true
    api_key_url: https://api.crowterminal.com/api/agent/register
requires:
  env:
    - CROWTERMINAL_API_KEY
Confidence
50% confidence
Finding
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

External Transmission

Medium
Category
Data Exfiltration
Content
clawdbot:
    category: productivity
    requires_api_key: true
    api_key_url: https://api.crowterminal.com/api/agent/register
requires:
  env:
    - CROWTERMINAL_API_KEY
Confidence
50% confidence
Finding
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

External Transmission

Medium
Category
Data Exfiltration
Content
clawdbot:
    category: productivity
    requires_api_key: true
    api_key_url: https://api.crowterminal.com/api/agent/register
requires:
  env:
    - CROWTERMINAL_API_KEY
Confidence
50% confidence
Finding
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

External Transmission

Medium
Category
Data Exfiltration
Content
clawdbot:
    category: productivity
    requires_api_key: true
    api_key_url: https://api.crowterminal.com/api/agent/register
requires:
  env:
    - CROWTERMINAL_API_KEY
Confidence
50% confidence
Finding
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

External Transmission

Medium
Category
Data Exfiltration
Content
clawdbot:
    category: productivity
    requires_api_key: true
    api_key_url: https://api.crowterminal.com/api/agent/register
requires:
  env:
    - CROWTERMINAL_API_KEY
Confidence
50% confidence
Finding
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

External Transmission

Medium
Category
Data Exfiltration
Content
clawdbot:
    category: productivity
    requires_api_key: true
    api_key_url: https://api.crowterminal.com/api/agent/register
requires:
  env:
    - CROWTERMINAL_API_KEY
Confidence
50% confidence
Finding
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

External Transmission

Medium
Category
Data Exfiltration
Content
### 1. Get API Key (Self-Registration)

```bash
curl -X POST "https://api.crowterminal.com/api/agent/register" \
  -H "Content-Type: application/json" \
  -d '{"agentName": "OpenClaw", "agentDescription": "My personal AI agent"}'
```
Confidence
60% confidence
Finding
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

External Transmission

Medium
Category
Data Exfiltration
Content
**Example: Natural language query**
```bash
curl -X POST "https://api.crowterminal.com/api/agent/memory/client_123/query" \
  -H "Authorization: Bearer $CROWTERMINAL_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"query": "engagement and hooks"}'
Confidence
60% confidence
Finding
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

External Transmission

Medium
Category
Data Exfiltration
Content
Send your current learnings, get back what configuration performed best:

```bash
curl -X POST "https://api.crowterminal.com/api/agent/memory/client_123/engagement-analysis" \
  -H "Authorization: Bearer $CROWTERMINAL_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
Confidence
88% confidence
Finding
This duplicate finding corresponds to the engagement-analysis submission of creator strategy data to an external processor. In the context of a persistent memory layer for influencers, transmitting optimization history and learned patterns can expose commercially sensitive information if users are not adequately warned.

External Transmission

Medium
Category
Data Exfiltration
Content
Send your current learnings, get back what configuration performed best:

```bash
curl -X POST "https://api.crowterminal.com/api/agent/memory/client_123/engagement-analysis" \
  -H "Authorization: Bearer $CROWTERMINAL_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
Confidence
88% confidence
Finding
This duplicate finding corresponds to the engagement-analysis submission of creator strategy data to an external processor. In the context of a persistent memory layer for influencers, transmitting optimization history and learned patterns can expose commercially sensitive information if users are not adequately warned.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The skill explicitly instructs users to upload creator/platform analytics data such as retention curves and watch time to a third-party service, but it provides no warning about privacy, consent, data classification, or downstream storage/processing risks. Because this skill is positioned as an AI memory layer for influencers/creators, the omitted warning materially increases the chance that sensitive or regulated business or personal data will be exfiltrated without informed review.

External Transmission

Medium
Category
Data Exfiltration
Content
Push platform data we can't access via API:

```bash
curl -X POST "https://api.crowterminal.com/api/agent/data/ingest" \
  -H "Authorization: Bearer $CROWTERMINAL_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
Confidence
96% confidence
Finding
The data-ingestion endpoint explicitly encourages sending platform data unavailable via public APIs, including retention curves and watch-time analytics, to a third-party service. Because this can include sensitive creator/business data and the skill omits cautionary language about privacy, authorization, and secure handling, it presents a genuine exfiltration and compliance risk.

Missing User Warnings

Medium
Confidence
92% confidence
Finding
The webhook section tells users to configure delivery of events to an arbitrary external URL but omits any warning that event payloads may contain creator or platform data and will be transmitted off-platform. This can lead users to forward sensitive operational data to insecure or unintended endpoints, especially if they copy the example without understanding what fields are included in webhook events.

External Transmission

Medium
Category
Data Exfiltration
Content
### Webhooks (Async Notifications)

```bash
curl -X POST "https://api.crowterminal.com/api/agent/webhooks" \
  -H "Authorization: Bearer $CROWTERMINAL_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
Confidence
91% confidence
Finding
The webhook registration example enables delivery of events to a user-specified URL but gives no indication of what data will be included or the security consequences of forwarding those events externally. This increases the likelihood of inadvertent leakage to poorly secured endpoints or attacker-controlled URLs if misconfigured.

Static analysis

No suspicious patterns detected.