Back to skill

Security audit

Contract And Tender Document Consistency Review Assistant 1.0.0

Security checks for vulnerabilities and agentic risk

Overview

This skill is a purpose-aligned contract and tender review helper, with disclosed document-handling integrations and no evidence of hidden code, persistence, exfiltration, or destructive behavior.

Install only if you intend to review Chinese procurement, bid, and contract materials. Provide documents deliberately, avoid unnecessary local-file access, and require a separate confirmation before allowing any optional tool to modify Office documents.

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
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (17)

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The README presents all user-facing instructions, trigger phrases, warnings, and usage guidance exclusively in Chinese. Under the policy rule, forcing a specific language without user opt-in or an explicitly documented locale constraint is a natural-language policy violation.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The trigger phrase "签约前核对" is generic and could match many ordinary pre-signing review requests, not specifically this bid-contract consistency skill. The manifest does not provide exclusion conditions or narrower scope constraints to distinguish this skill from other contract-checking or document-review skills.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
87% confidence
Finding

The single-word trigger “阴阳合同” is broad enough to fire on casual discussion, legal education, or news analysis, causing unintended invocation of a skill that may inspect sensitive local procurement documents. In this skill’s context, accidental activation matters more because it is wired to document-reading and optional document-editing capabilities.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
87% confidence
Finding

The trigger “黑白合同” has the same ambiguity problem as other single-term legal phrases and may activate during ordinary discussion rather than a deliberate request to run the skill. Because the skill can route into local document handling, unintended activation could expose confidential bid, contract, or award materials to unnecessary processing.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The manifest declares optional local document reading and editing capabilities but does not warn users that the skill may access or modify local Office/PDF files. In a procurement-contract review workflow, those files are often sensitive and high-value, so lack of disclosure and confirmation increases the risk of unintended data exposure or accidental document alteration.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
84% confidence
Finding

The trigger "contract deviates from tender" is descriptive but still broad enough to overlap with general legal or procurement discussion, rather than a clear invocation phrase for this specific skill. No negative examples or invocation constraints are provided in the manifest to limit accidental matches.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The repeated trigger phrase "签约前核对" remains generic in the skillhub platform section and could match many routine pre-signing checks outside this skill's intended tender/bid consistency context. The manifest still lacks narrowing conditions or negative examples.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
87% confidence
Finding

The repeated single-term trigger “阴阳合同” is still ambiguous in this platform section and can cause unintended invocation in contexts unrelated to document review. Since the skill may connect to local document tooling and knowledge bases, accidental activation could pull in sensitive contract materials without sufficient user intent verification.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
87% confidence
Finding

The repeated “黑白合同” trigger is ambiguous and may match ordinary legal discussion, increasing the chance of unintended skill execution. In this skill, the consequence is not arbitrary code execution, but unnecessary handling of confidential bidding and contract documents, which is still a meaningful security/privacy issue.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The repeated trigger phrase "签约前核对" in the cocoloop section remains overly general and not uniquely tied to this consistency-review skill. Without constraints, it may catch ordinary pre-signature review requests.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
87% confidence
Finding

The cocoloop-section trigger “阴阳合同” remains an ambiguous single-term activator that can match discussions not intended to invoke document analysis. Given the skill’s access pattern to sensitive procurement materials, such ambiguity raises the risk of privacy-impacting misfires.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
87% confidence
Finding

The cocoloop-section trigger “黑白合同” is likewise too short and context-poor, increasing accidental activation risk. Because this skill’s workflow may involve confidential contracts and potentially file-editing integrations, even a non-malicious misfire can have material data-handling consequences.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The repeated trigger phrase "签约前核对" in the coze section is still too generic and likely to overlap with everyday legal or business review requests. The manifest provides no explicit trigger constraints or non-trigger examples.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
87% confidence
Finding

The coze-section single-term trigger “阴阳合同” can activate on broad legal conversation and is insufficiently precise for a skill with potential access to local files and sensitive procurement content. This makes unintended processing of confidential materials more plausible.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
87% confidence
Finding

The coze-section trigger “黑白合同” is similarly ambiguous and may cause accidental invocation outside intended contract-consistency review workflows. In context, the danger is amplified by the skill’s stated integrations for reading and potentially editing local documents.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

This markdown file contains all user-facing natural-language instructions and test criteria in Chinese only. Under the policy, forcing a specific language without user opt-in or a clearly justified locale constraint is a natural-language policy violation.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
87% confidence
Finding

The instruction requires the assistant to use a fixed style and terminology in Chinese: "语气客观、冷峻、权威" and specific Chinese legal terms. This imposes a language/locale preference without any user opt-in or alternative-language option, which can violate language-choice policy for general-use skills.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.