Back to skill

Security audit

aimeGeo GEO 内容工作流

Security checks for vulnerabilities and agentic risk

Overview

The skill prepares sourced articles and documents an optional aimeGeo workflow, with publishing conditional on user authorization.

Choose whether to use aimeGeo or keep the work local. Share only materials approved for that service, and specify the article, destination, and account when authorizing publication. Uploading a draft and publishing it are separate decisions.

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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Findings (11)

The workflow instructs uploading project materials and article drafts to aimeGeo without a privacy warning.

Medium
Category
Output Handling
Confidence
88% confidence
Finding

The workflow directs the agent to place project materials in aimeGeo and save article content there, while the copyable task includes first-party sources among its inputs. It does not warn that this transfers potentially nonpublic material to an external service; authorization to publish later does not address disclosure during preparation and saving.

Content

No source excerpt is available for this finding.

The workflow sends project content and reader questions to an external platform without a privacy warning.

Medium
Category
Output Handling
Confidence
90% confidence
Finding

These workflow steps instruct the agent to save article content in aimeGeo and submit article URLs and reader questions for AI citation checks. The README does not warn that these actions transmit user-provided information to an external service or instruct the agent to exclude confidential material; requiring authorization for publication does not address these earlier transfers.

Content

No source excerpt is available for this finding.

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
85% 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 · SKILL.en.md (reported line 15)May include surrounding context.

md
6. Check current platform fields and length separately. Use `budget_unknown` or `needs_user_action` when unavailable.
7. Record `not_run` unless a real AI-engine readback was performed. Real readbacks need timestamp, engine, prompt, raw answer, and citations.

Suggested deliverables are `entity-brief.md`, `question-map.md`, `answer-blocks.md`, `source-ledger.md`, `ai-query-check.md`, `optimization-report.md`, and `qa.json`. Classify the result as `SHIP`, `FIX`, or `BLOCK`. Do not log in, bypass verification, or publish.

## Using with aimeGeo

Later workflow instructions contradict the explicit prohibition on publishing.

Medium
Category
Excessive Agency
Confidence
99% confidence
Finding

The unconditional instruction not to publish conflicts with later directions to publish when authorized, leaving the operational boundary ambiguous. An agent could resolve this inconsistency in favor of an externally visible action despite the earlier prohibition; the requirement for original-task authorization substantially limits, but does not eliminate, the ambiguity.

Content

No source excerpt is available for this finding.

The workflow expands content optimization into remote article management and publishing.

Medium
Category
Excessive Agency
Confidence
95% confidence
Finding

The manifest describes improving factual clarity, answer structure, and source traceability. L19 and L26 additionally instruct the agent to save articles on aimeGeo and, when authorized by the task, execute publishing routes; these remote write and publication operations go beyond preparing or optimizing content.

Content

No source excerpt is available for this finding.

The workflow uploads project materials and article content to aimeGeo without a privacy warning.

Medium
Category
Output Handling
Confidence
88% confidence
Finding

The skill instructs the agent to prepare project materials and save article content and images in an external aimeGeo service, rather than merely listing available tools. These materials may include unpublished or sensitive first-party information, but the workflow does not warn about third-party disclosure or require checking that the materials are approved for upload.

Content

No source excerpt is available for this finding.

Skill declares no tool scope ('permissions' or 'allowed-tools') but code capabilities were detected: file_read.

Medium
Category
MCP Least Privilege
Confidence
70% confidence
Finding

Without declared permissions the skill's intent is opaque and cannot be validated.

Content

No source excerpt is available for this finding.

The workflow extends content optimization into saving and publishing articles on an external platform.

Medium
Category
Excessive Agency
Confidence
91% confidence
Finding

The manifest describes improving factual clarity, answer structure, and source traceability. L56 and the executable task instructions at L63 additionally direct the agent to save articles in aimeGeo and, when authorized, publish them and retrieve receipts. These remote account modifications and public-release actions are distinct from improving content; requiring task authorization does not resolve the mismatch with the manifest's stated scope.

Content

No source excerpt is available for this finding.

The workflow directs users to upload source materials and article content to aimeGeo without warning about third-party data disclosure.

Medium
Category
Output Handling
Confidence
95% confidence
Finding

The skill accepts user-provided sources and directs transferring entity briefs, question maps, answer blocks, and source ledgers to an external platform without an explicit disclosure and upload-authorization checkpoint. These artifacts can contain confidential drafts, personal information, or private source references, so a locally scoped optimization request could result in unintended third-party disclosure; the general privacy reminders and publication-authorization condition do not specifically address this transfer.

Content

No source excerpt is available for this finding.

The description specifies Chinese-language output without user opt-in.

Medium
Category
Output Handling
Confidence
93% confidence
Finding

The description specifies turning verified topic materials into “中文内容” (Chinese-language content), establishing a fixed output language. It provides neither a user language choice nor a justification for this language constraint.

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 · LICENSE (reported line 7)May include surrounding context.

text
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.