Report Generator EN

• Outcome: Pass
• Latest audit: Sep 15, 2026 · 1:54 AM UTC
• Version: 1.0.1

Overview

This is a prompt-only weekly report generator with no executable behavior, but users should avoid pasting confidential work, HR, finance, customer, or incident details into third-party AI chats.

Install/use is reasonable for ordinary report drafting. Before using it with real workplace material, redact secrets, customer data, employee personal data, incident details, unreleased financials, credentials, and other confidential information unless your organization approves the AI provider and data handling path. The publisher should also fix the remaining zero-width joiner and update the changelog accurately.

A.I.G

Vulnerability Patterns

• Insecure Skill Coding Practices: Finds exploitable flaws such as hardcoded secrets or command injection
• Skill Instruction Hijacking: Alters the agent's session goals or safety constraints when the skill loads
• Agent Memory Poisoning: Writes attacker-controlled rules into memory that affect later sessions
• Remote Payload Retrieval and Execution: Fetches external code whose behavior can change after review
• Embedded Malicious Code: Ships malicious scripts inside the skill and executes them locally

Findings (1)

T09 · Insecure Skill Coding Practices

Note

• Location: SKILL.md:354

Finding

Undocumented Zero-Width Joiner Remains in the Skill Prompt

Content

View full analysis

Vulnerability Details

File Location: SKILL.md, line 354
Vulnerability Type: Invisible Unicode character in security-sensitive prompt text
Risk Level: Low

markdown

### 🧑‍💼 HR Weekly Report

The emoji sequence contains U+200D ZERO WIDTH JOINER between its visible components. This contradicts the statement in CHANGELOG.md, lines 7–10, that the character was removed and the emoji was replaced with 👤.

Technical Analysis

SKILL.md is intended to be loaded directly as an AI system prompt. Invisible Unicode formatting characters can create differences between the text displayed to reviewers and the underlying character sequence processed by scanners or downstream systems. Such discrepancies can weaken source review, normalization, pattern matching, and prompt-integrity checks.

In this specific instance, U+200D only combines emoji components. The audited text contains no hidden command, instruction override, or executable payload associated with the character. Consequently, the finding represents an insecure prompt-maintenance and validation practice rather than demonstrated prompt injection.

Attack Path

1. A user follows the documented instructions in README.md and loads the complete SKILL.md file as a system prompt.
2. The U+200D character at line 354 enters the agent instruction context as part of the prompt.
3. A scanner or review process that normalizes, ignores, or renders invisible Unicode differently may evaluate a representation that differs from the actual prompt.
4. If similar characters were later used around security-sensitive instructions, an attacker could exploit that review discrepancy to obscure prompt modifications.

The current character does not independently alter agent behavior, grant privileges, or execute code. No complete malicious exploitation path is present in the audited project.

Impact Assessment

The immediate impact is limited to prompt integrity, audit reliability, and release-verif ...[truncated 339 chars]

Remediation

View remediation

Remediation Suggestions

1. Replace 🧑‍💼 with the single-code-point symbol 👤, as stated in CHANGELOG.md.
2. Normalize prompt and documentation files to a defined Unicode normalization form during the release process.
3. Add a CI check that rejects unexpected zero-width and bidirectional control characters, including U+200B, U+200C, U+200D, U+2060, U+FEFF, and U+202A–U+202E.
4. Permit such characters only through an explicit allowlist when they are necessary and documented.
5. Compare normalized and raw file representations during security review to identify rendering discrepancies.
6. Verify source changes before publishing security claims in the changelog.

SkillSpector

Vulnerability Patterns

• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (3)

Missing User Warnings

Medium

• Category: Not specified by scanner
• Confidence: 91% confidence

Finding

The README instructs users to paste weekly work details into third-party AI chat tools but does not warn against including confidential business, customer, personnel, or security-sensitive information. Because weekly reports often contain internal metrics, incidents, project status, and staffing details, this omission can lead to unintentional data disclosure to external AI providers or unauthorized systems.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

• Category: Not specified by scanner
• Confidence: 96% confidence

Finding

The instruction states that the default language is English unless the user explicitly asks otherwise. This imposes a specific language preference by default rather than offering a neutral language choice or explicit opt-in, which matches the language/locale policy violation criteria.

Content

No source excerpt is available for this finding.

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 · SKILL.md (reported line 11)May include surrounding context.

md

## Skill Definition

You are a professional weekly-report writing assistant. Your job is to turn scattered work updates into clear, professional, insight-driven, structured weekly reports. You support a wide range of roles including (but not limited to): frontend engineer, backend engineer, mobile engineer, QA engineer, operations, sales, HR, accounting, product manager, designer, marketing, etc.

### Core Capabilities