Back to skill

Security audit

Pogo Pvp

Security checks for vulnerabilities and agentic risk

Overview

The skill is a coherent Pokémon GO PvP assistant, but it should be reviewed because its main skill description under-discloses local inventory writes and one evaluation command silently changes saved data.

Install only if you are comfortable with a Chinese-language Pokémon GO PvP assistant that stores and updates a local roster file. Before using it, review data/my_pokemon.json behavior, consider making backups, and prefer changing /pvp 评估我的 so it reports recommended moves without saving unless you explicitly confirm.

Vulnerability Patterns
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • 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
Findings (1)

T09 · Insecure Skill Coding Practices

Warning
Location
src/evaluate.ts:235
Finding

Unvalidated Remote Ranking Data Silently Modifies User Inventory

Content
View full analysis
x.speciesId === p.speciesId && x.league === p.league && x.iv[0] === p.iv[0] && x.iv[1] === p.iv[1] && x.iv[2] === p.iv[2] ); if (idx >= 0) { allP[idx].moves = movesToSave; allP[idx].updatedAt = new Date().toISOString(); fs.writeFileSync(DATA_FILE, JSON.stringify(allP, null, 2), 'utf-8'); } } ``` ### Technical Analysis The application downloads ranking and gamemaster JSON from fixed GitHub Raw URLs and accepts each response solely on the basis that it is syntactically valid JSON. The TypeScript assertion `as PvPokePokemon[]` does not perform runtime validation. There are no checks for the expected top-level structure, required properties, value types, numeric ranges, string lengths, allowable move identifiers, ...[truncated 2413 chars]
Remediation
View remediation
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Findings (38)

Tp4

High
Category
MCP Tool Poisoning
Confidence
97% confidence
Finding

The declared description emphasizes query/analysis capabilities: ranking lookup, best IV, and IV comparison based on PvPokeTW and gamemaster data. The supplied code chunk does not perform ranking lookup or IV analysis. Instead, its primary function is CRUD-like data persistence for a personal Pokémon list: it parses input, validates fields, resolves a species ID, loads and saves data/my_pokemon.json, and updates existing entries. This is a materially different capability from the declared purpose and also involves filesystem access not reflected in the declared permissions. While such storage could support a broader PvP tool, this chunk itself implements undeclared add/update persistence behavior rather than the described query features.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
96% confidence
Finding

声明描述聚焦于“培养查询”、真实排行数据、最佳 IV 和用户 IV 对比,这与代码中的一部分功能相符,例如 queryPokemon、queryIV 和结果格式化。但实际代码显著扩展到了多种未在描述中体现的能力:1) 个人数据管理与评估(handleAdd、handleList、handleEvaluate);2) 配队建议/构建(buildTeam);3) 基于已有收藏进行培养优先级、缺口分析和推荐(buildTrainList、buildOrderList、findMissing);4) 除技能命令触发外还提供了命令行执行入口。由于这些能力不是单纯的支撑实现细节,而是面向用户的独立主要功能,实际行为比声明范围更广,描述不能准确完整代表代码行为,因此应判定为不匹配。

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

声明描述的核心能力是 PvP 培养查询:排名查询、最佳 IV、用户 IV 对比,重点在个体值与培养决策。实际代码没有任何 IV 排名、IV 比较、个体值输入处理或最佳 IV 计算逻辑;相反,它完整实现了另一个主要功能——队伍构建与评分。代码会加载 PvP 排名与 gamemaster,读取用户持有宝可梦列表,按联盟和可选核心宝可梦生成三人队伍,推荐招式,并基于属性克制计算环境覆盖、战术弹性、队伍容错和关键威胁。这属于 materially different primary purpose,因此应判定为描述与行为不匹配。

Content

No source excerpt is available for this finding.

Vague Triggers

High
Category
Not specified by scanner
Confidence
93% confidence
Finding

A single-character noun like "龟" is extremely ambiguous and can appear in many unrelated contexts. If this manifest participates in activation or routing, this creates a high risk of accidental matches and unintended skill invocation.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
70% confidence
Finding

npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
70% confidence
Finding

npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
78% confidence
Finding

The skill documentation explicitly describes network retrieval of external ranking and gamemaster data, but the manifest shown in SKILL.md does not declare any tool scope or permissions. Undeclared network capability weakens least-privilege controls and can hide data flows or future expansion of external requests from reviewers and users.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The description states the skill is a Chinese Pokémon GO PvP query skill, and the command/interface examples throughout the file are presented only in Chinese. There is no indication that users can opt into another language or that the Chinese-only constraint is a documented, justified regional requirement.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The example output and the surrounding document consistently prescribe Chinese labels and content, but there is no indication that users may choose another language or locale. This can violate a language/locale policy when the skill behavior is effectively fixed to a specific language without explicit user opt-in.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The design states '禁止输出纯英文招式名' and requires outputs to use Chinese whenever available, enforcing a specific language/locale behavior by default. This is a natural-language policy concern because the document does not mention any user choice, opt-in, or locale selection.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The design explicitly states that all Pokémon names in output must use Chinese, which imposes a fixed language choice. This is a natural-language policy concern because the file does not indicate any user language preference, opt-in, or fallback behavior.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
87% confidence
Finding

This JSON manifest contains very generic terms such as "水母" that commonly appear in ordinary conversation and are not specific enough to reliably indicate intent to invoke a Pokémon-related skill. Similar broad aliases elsewhere in the file increase the risk of unintended activation if these values are used as trigger phrases.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

"月亮" is a common everyday noun and does not clearly indicate the skill should activate for Cresselia-related behavior. Without scope limits or exclusions, such broad natural-language aliases can collide with normal user speech.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

"太阳" is a highly common term unrelated to Pokémon in many contexts, making it too vague if interpreted as a trigger or activation phrase. The file provides no explicit constraints distinguishing lookup-only aliases from invocation terms.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The manifest description embeds very broad trigger phrases such as "PvP" and "宝可梦对战", which are generic terms likely to appear in many normal conversations about Pokémon GO. This can cause the skill to activate unintentionally, creating routing confusion and allowing the skill to handle requests outside the user's clear intent.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

This code persistently stores user-provided Pokémon data to a local JSON file, which goes beyond a purely query-oriented skill and creates retained state containing user input. Persistent storage increases privacy and integrity risk because data remains on disk, can be unintentionally exposed to other components or users, and changes the trust model from read-only lookup to stateful collection.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The add command inserts and updates records in local storage rather than only performing lookup or comparison, so the skill can mutate host state based on user input. In a query-focused skill, this is a capability expansion that may surprise operators, enable unauthorized data modification, and create integrity issues if the environment assumes the skill is read-only.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The file’s natural-language comments and all returned user-facing messages are written exclusively in Chinese, and the command description /pvp 培养顺序 implies a fixed-language interaction. There is no indication that users can choose a different language or that the locale restriction is documented as a justified region-specific constraint.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The /pvp 评估我的 operation is presented as a read/evaluate command, but it mutates persistent state by auto-filling p.moves and writing my_pokemon.json back to disk. Hidden writes in a query path are dangerous because they violate user expectations, can overwrite intentionally blank fields, and create integrity/audit problems if recommendations are wrong or stale.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The function silently persists recommended move data to the user's JSON file with no warning, confirmation, or indication in the returned output. In this skill context, the risk is primarily integrity and transparency: users think they are performing an evaluation, but the command changes stored inventory data, which can cascade into later decisions based on modified records.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The manifest describes this skill as providing ranking queries, best IV, and user-IV comparison based on PvP data. This entrypoint also exposes add/list/evaluate, team building, training priority, build order, and missing-roster analysis commands, which are materially broader than simple lookup/query functionality.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The docstring specifies a Chinese Pokémon name input and the file’s user-facing messages and labels are entirely in Chinese. Under the policy, forcing a specific language without user opt-in is a natural-language policy violation unless the locale constraint is explicitly documented and justified as region-specific, which is not evident here.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The file comments and exported mappings are explicitly dedicated to Chinese names only, and the public helper functions always return Chinese labels when a mapping exists. This indicates a hard-coded language choice with no visible mechanism for user opt-in or locale selection, which matches the language/locale policy violation criteria.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

This file contains user-facing natural language entirely in Chinese, including the command description and all returned messages, with no indication that the user can opt into this language or that the skill is intentionally limited to a Chinese-speaking audience. That can violate a language/locale policy requiring user choice or explicit justification for locale-specific behavior.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

This file returns user-facing text entirely in Chinese, including usage guidance and error messages, which forces a specific language on users. The policy allows locale constraints only when they are explicitly offered as a choice or clearly documented and justified as region-specific.

Content

No source excerpt is available for this finding.