Back to skill

Security audit

Developer Docs Framework

Security checks for vulnerabilities and agentic risk

Overview

This is a documentation-guidance skill with broad triggers and some credential-example guidance to review, but its artifacts do not show hidden execution, persistence, or data theft.

Install this if you want broad documentation-framework help. Review whether you want such broad docs-related triggers, and do not adopt the Stripe-style actual-test-key personalization pattern unless your docs platform masks, scopes, and controls credentials safely.

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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (19)

Hidden Instructions

High
Category
Prompt Injection
Content
2. Copy the structure into your document
3. Replace placeholder text (marked with `[brackets]`) with your content
4. Remove any optional sections that don't apply
5. Follow the inline guidance comments (marked with `<!-- -->`)

---
Confidence
70% confidence
Finding
Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.

Hidden Instructions

High
Category
Prompt Injection
Content
## Configure

<!-- Show the minimum configuration needed. -->

```[language]
[minimum configuration code]
Confidence
70% confidence
Finding
Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.

Hidden Instructions

High
Category
Prompt Injection
Content
## Verify

<!-- How to confirm the task completed successfully -->

[Verification steps or expected outcome]
Confidence
70% confidence
Finding
Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.

Hidden Instructions

High
Category
Prompt Injection
Content
## Authentication

<!-- How to authenticate requests between systems -->

[Authentication setup steps with code]
Confidence
70% confidence
Finding
Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.

Hidden Instructions

High
Category
Prompt Injection
Content
## Step 1: [First integration step]

<!-- Show both sides of the integration where relevant -->

[Steps with code examples]
Confidence
70% confidence
Finding
Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.

Hidden Instructions

High
Category
Prompt Injection
Content
## Automated migration

<!-- If available, show automated migration tools first -->

```shell
[codemod or migration script command]
Confidence
70% confidence
Finding
Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.

Hidden Instructions

High
Category
Prompt Injection
Content
```markdown
## [Method] [Path]

<!-- e.g., POST /v1/payments -->

[One sentence: what this endpoint does]
Confidence
70% confidence
Finding
Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.

Hidden Instructions

High
Category
Prompt Injection
Content
## [Version] — [Date]

<!-- One sentence summary of the most significant change -->

### Added
- [New feature with link to documentation]
Confidence
70% confidence
Finding
Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.

Vague Triggers

Medium
Confidence
93% confidence
Finding
The 'When to Use This Skill' section says the skill is for 'writing, planning, auditing, and improving technical documentation,' which is very broad and overlaps with common documentation requests. It does not provide explicit trigger phrases, scope limits, or negative examples to distinguish when this skill should or should not activate.

Natural-Language Policy Violations

Medium
Confidence
88% confidence
Finding
The file prescribes English-language stylistic rules such as addressing the reader as 'you,' using active voice, and spelling out acronyms, and later references global readability in ways that assume English as the output language. Because the skill does not clearly offer language or locale choice up front, these defaults can conflict with a policy requiring user opt-in before forcing a specific language or locale.

Vague Triggers

Medium
Confidence
96% confidence
Finding
The skill declares a very broad set of trigger phrases such as "write docs," "document this," "tutorial," and "getting started," which are common in normal developer workflows and can cause the skill to activate in many unrelated contexts. Over-broad activation increases the chance of unintended instruction injection into conversations, crowding out more appropriate skills and making agent behavior less predictable.

Missing User Warnings

Medium
Confidence
88% confidence
Finding
The file recommends that documentation examples use a developer's actual test API keys when logged in, which normalizes auto-inserting credentials into rendered docs. Even though these are test keys, exposing or prepopulating secrets in browser-visible examples increases the chance of credential leakage through screenshots, shared sessions, browser extensions, logs, or accidental reuse patterns that later affect live credentials. In a documentation-framework skill, this is more dangerous because it may be adopted as a best practice across many downstream docs implementations.

External Transmission

Medium
Category
Data Exfiltration
Content
## 1. Register your endpoint (your side → our API)

```bash
curl -X POST https://api.example.com/v1/telemetry/subscribe \
  -d callback_url="https://partner.com/telemetry" \
  -d device_types[]="sensor.temperature"
```
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
## 1. Register your endpoint (your side → our API)

```bash
curl -X POST https://api.example.com/v1/telemetry/subscribe \
  -d callback_url="https://partner.com/telemetry" \
  -d device_types[]="sensor.temperature"
```
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
## 1. Register your endpoint (your side → our API)

```bash
curl -X POST https://api.example.com/v1/telemetry/subscribe \
  -d callback_url="https://partner.com/telemetry" \
  -d device_types[]="sensor.temperature"
```
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. Register your endpoint (your side → our API)

```bash
curl -X POST https://api.example.com/v1/telemetry/subscribe \
  -d callback_url="https://partner.com/telemetry" \
  -d device_types[]="sensor.temperature"
```
Confidence
50% confidence
Finding
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Natural-Language Policy Violations

Low
Confidence
86% confidence
Finding
The document explicitly states 'Always use "you"' and presents it as a blanket override for all tutorials and related content. This enforces a specific language-style policy without requiring user choice or limiting it to clearly justified regional or compliance needs.

Natural-Language Policy Violations

Low
Confidence
79% confidence
Finding
The checklist item 'Localization for key markets' prescribes locale-specific behavior as a maturity requirement, but it does not mention user opt-in, choice, or justification for which locales are prioritized. This can conflict with a language/locale policy that requires explicit choice or documented justification rather than imposing localization direction unconditionally.

Missing User Warnings

Low
Confidence
81% confidence
Finding
This markdown file contains interactive example snippets that show `api_key="cs_test_..."` and similar client initialization tokens. Because the file does not warn readers that these are placeholder test credentials, users may misunderstand the examples and handle credentials unsafely when copying the snippet.

Static analysis

No suspicious patterns detected.