Security audit

AI legal assistant for Indian advocates and lawyers

Security checks for vulnerabilities and agentic risk

Overview

This is a coherent India-focused legal drafting and case-analysis skill, with sensitive legal data risks that are disclosed enough for normal cautious use.

Install only if you want an India-focused legal assistant for advocate workflows. Use redacted placeholders where possible, avoid pasting full government IDs or bank details unless strictly needed, and have all legal drafts and citations checked by a qualified advocate before filing.

A.I.G

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

SkillSpector

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 (8)

Vague Triggers

High

Category: Not specified by scanner
Confidence: 97% confidence
Finding: The trigger scope is excessively broad and includes common phrases such as "my client," "HC," "SC," "matter," and "brief," which can cause the skill to activate in many unrelated conversations. In a legal skill that may read uploaded case files and handle sensitive client data, unintended activation increases the risk of inappropriate context capture, privacy exposure, and user confusion about when specialized instructions are being applied.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 88% confidence
Finding: The instruction "Use Indian legal English" imposes a specific language/locale requirement as a default behavior. The file does not indicate that the user can choose another language or opt into this locale-specific output format.
Content: No source excerpt is available for this finding.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 95% confidence
Finding: The guidance instructs analysis of uploaded legal documents but omits any warning about confidential, privileged, personal, or criminal-case data. Because legal documents routinely contain addresses, IDs, witness details, medical facts, and attorney-client material, failing to warn or gate handling can lead to privacy breaches and inappropriate disclosure or retention of sensitive information.
Content: No source excerpt is available for this finding.

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 91% confidence
Finding: The workflow is triggered for essentially any uploaded legal document, which can cause the skill to process highly sensitive case files, FIRs, judgments, and client records without a narrower activation boundary or explicit user confirmation. In a legal-assistant context this increases the chance of unnecessary exposure of privileged, personal, and criminal-justice data to the skill when the user may not intend full document analysis.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 84% confidence
Finding: The guidance is written around Indian legal frameworks and cites Indian statutes and procedures, but the file does not clearly state up front that the skill is limited to Indian-law matters. This can create a locale-policy issue because it assumes a specific legal jurisdiction without offering user choice or an explicit documented scope boundary.
Content: No source excerpt is available for this finding.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 91% confidence
Finding: The affidavit and related legal templates instruct users to include sensitive personal data such as full name, age, occupation, residential address, and case-identifying facts, but provide no warning to minimize, redact, or securely handle that data. In a legal-assistant skill, this increases the likelihood that users will paste highly sensitive client and case information into prompts or generated documents without safeguards, creating privacy and confidentiality risk.
Content: No source excerpt is available for this finding.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 94% confidence
Finding: The intake form explicitly requests highly sensitive identifiers such as Aadhaar and PAN in a general template, while giving only a minimal note to redact Aadhaar in files and store it securely. In a legal-assistant skill that may process uploaded documents and client intake data, this increases the chance of unnecessary collection, over-retention, or accidental disclosure of regulated personal data.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Low

Category: Not specified by scanner
Confidence: 76% confidence
Finding: The reference uses India-specific legal and administrative concepts such as Aadhaar, PAN, GST thresholds, and Indian limitation-law references, but it does not explicitly state that the guidance is intended for Indian legal practice. This can create a locale-policy issue because users are not given a choice or clear notice that the content is jurisdiction-specific.
Content: No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.