Back to skill

Security audit

visitor-identification

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed Cargo/Snitcher visitor-identification setup that waits for visitor consent and requires explicit approval before deployment, though it can incur tracking costs.

Before installing, confirm your team is allowed to use Snitcher/Cargo visitor identification, provide a reviewed privacy disclosure, check the live per-company price and credit balance, and only approve deployment after reviewing the plan and site consent behavior.

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
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (10)

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 90)May include surrounding context.

md
5. **Plan, then stop.** `node --import tsx evals/contract.mjs` from this skill's folder, then

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 113)May include surrounding context.

md
5. **Plan, then stop.** `node --import tsx evals/contract.mjs` from this skill's folder, then

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 147)May include surrounding context.

md
5. **Plan, then stop.** `node --import tsx evals/contract.mjs` from this skill's folder, then

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 132)May include surrounding context.

md
- **Nothing loads before an explicit accept.** (`site/components/visitor-consent.tsx`) Snitcher's

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
80% confidence
Finding

The skill describes actions that read environment-provided configuration and interact with deployment/connection state, but it does not declare any explicit tool scope or permissions boundary. In an agent setting, missing scope declarations can allow broader-than-expected access or execution paths, making it easier for the skill to invoke sensitive capabilities without clear review.

Content

No source excerpt is available for this finding.

Whitespace Padding

Medium
Category
Prompt Injection
Confidence
70% confidence
Finding

Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.

Content

Scanner excerpt · SKILL.md (reported line 101)May include surrounding context.

md
**Derive before you ask.** An input with a lookup is looked up, not asked.

| Input                                                    | Kind    | How it is answered                                                                                         | Why it matters                                                                                                          |
| -------------------------------------------------------- | ------- | ---------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------- |
| site URL (`infra/models/companies.ts`)                   | derived | the website app's `site.json` `canonicalUrl`, which must be HTTPS and the site `ready`                     | Snitcher provisions a workspace for this exact origin, and the gate only runs on it. Changing it later starts over.    |
| Snitcher connection (`infra/connectors/snitcher.ts`)     | derived | `cargo-ai connection connector list`: the workspace's default Snitcher connection, on Cargo's credits      | The managed path provisions the tracker. A workspace's own Snitcher key provisions nothing and needs a different setup. |

Whitespace Padding

Medium
Category
Prompt Injection
Confidence
70% confidence
Finding

Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.

Content

Scanner excerpt · SKILL.md (reported line 120)May include surrounding context.

md
**Derive before you ask.** An input with a lookup is looked up, not asked.

| Input                                                    | Kind    | How it is answered                                                                                         | Why it matters                                                                                                          |
| -------------------------------------------------------- | ------- | ---------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------- |
| site URL (`infra/models/companies.ts`)                   | derived | the website app's `site.json` `canonicalUrl`, which must be HTTPS and the site `ready`                     | Snitcher provisions a workspace for this exact origin, and the gate only runs on it. Changing it later starts over.    |
| Snitcher connection (`infra/connectors/snitcher.ts`)     | derived | `cargo-ai connection connector list`: the workspace's default Snitcher connection, on Cargo's credits      | The managed path provisions the tracker. A workspace's own Snitcher key provisions nothing and needs a different setup. |

Whitespace Padding

Medium
Category
Prompt Injection
Confidence
70% confidence
Finding

Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.

Content

Scanner excerpt · SKILL.md (reported line 101)May include surrounding context.

md
**Derive before you ask.** An input with a lookup is looked up, not asked.

| Input                                                    | Kind    | How it is answered                                                                                         | Why it matters                                                                                                          |
| -------------------------------------------------------- | ------- | ---------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------- |
| site URL (`infra/models/companies.ts`)                   | derived | the website app's `site.json` `canonicalUrl`, which must be HTTPS and the site `ready`                     | Snitcher provisions a workspace for this exact origin, and the gate only runs on it. Changing it later starts over.    |
| Snitcher connection (`infra/connectors/snitcher.ts`)     | derived | `cargo-ai connection connector list`: the workspace's default Snitcher connection, on Cargo's credits      | The managed path provisions the tracker. A workspace's own Snitcher key provisions nothing and needs a different setup. |

Whitespace Padding

Medium
Category
Prompt Injection
Confidence
70% confidence
Finding

Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.

Content

Scanner excerpt · SKILL.md (reported line 120)May include surrounding context.

md
The code is a worked example. These reshapes are expected, and the agent offers them rather than
waiting to be asked. Every one costs something.

| Variation         | When it is right                                          | How                                                                                                                   | What it costs                                                                                         |
| ----------------- | --------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------- |
| `existing-cmp`    | The site already runs a consent management platform       | Call `loadSnitcher` from the CMP's grant event and its deny path from revocation, instead of rendering the gate      | The same behaviour has to be re-tested against the CMP's events, including withdrawal.                |
| `scoring-input`   | Visits should move an account's score                     | Read the companies model from `account-scoring` and weight recency and pages read                                     | A score that leans on visits favours companies with more staff online, not more intent.               |

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 · site/lib/visitors.ts (reported line 11)May include surrounding context.

ts
privacyPolicyUrl: string;
};

// Read at build time by app/layout.tsx. `null` renders no consent gate, so a
// draft site, a preview, or a build without the snippet loads nothing: there
// is no profile ID in the export to load it with.
//