Security audit
Security checks for vulnerabilities and agentic risk
This is a Chinese I Ching divination skill that runs local hexagram and calendar calculations without persistence, credential access, or network behavior.
Install only if you want a Chinese-language divination and symbolic-analysis skill. Be aware that its trigger phrases are broad and that it may proceed with casting after a brief confirmation; avoid using it for medical, legal, financial, or other consequential decisions as a source of certainty.
The changelog explicitly states that any user response is treated as a trigger to initiate divination, which removes meaningful user consent and makes activation overly broad. In an agent context, this can cause unintended execution of the skill on casual or unrelated input, increasing the risk of surprise behavior, privacy issues, and misuse of user-provided text as operational input.
The description promises a divination/game capability centered on interpreting six-line hexagram patterns to guess a concealed object. The code does not accept or analyze hexagram lines, does not infer any object, and does not implement《正易心法》卦脉法. Its sole function is calendrical support: deriving year干支, month建, month干支, day干支, and five-element attributes from the current date/time using simplified formulas and solar-term boundaries. While such date calculations could be a supporting component for some I Ching-related systems, this chunk’s actual primary purpose is clearly a历法 module, which is materially different from the declared end-user capability.
The declared purpose focuses on 射覆: analyzing an already-formed six-line structure to infer the concealed object via a specific interpretive method. This code instead performs the upstream step of 起卦 using the 大衍筮法, generating random lines through simulated yarrow-stalk operations, then mapping line arrays to hexagram names via a database. It outputs 本卦、变卦、动爻 and calendar reference data, but contains no logic for guessing or describing a concealed object, no 《正易心法》卦脉 reading logic, and no direct interpretive layer. The database lookup also conflicts with the claim of '不查表'. Therefore the code's actual primary behavior materially differs from the declared description.
The declared description emphasizes a specific divination use case: 射覆, i.e. guessing a concealed object from the hexagram via《正易心法》卦脉法 and direct qi interpretation. However, the supplied code only performs hexagram generation and conventional 六爻装卦 steps. It tosses coins or parses manual toss input, loads gua and 纳甲 data from JSON files, fetches Ganzhi calendar info from a helper module, calculates 六亲 and 六神, identifies 世应, and returns a structured representation of the base and changing hexagrams. There is no logic that analyzes the hexagram to predict a hidden object, no 射覆-specific inference engine, and no implementation of the claimed 'direct reading' method. In fact, the code explicitly depends on lookup tables, which conflicts with the description's '不查表' claim. Therefore the behavior materially differs from the declared purpose.
The declared description emphasizes a specific use case and method: 射覆 (guessing a concealed object) via 《正易心法》卦脉法, directly reading qi trends from six lines and explicitly not using lookup tables or stock formulas. The actual code does something materially different: it performs 梅花易数起卦, deriving upper/lower trigram and moving line from time or numeric input, loading gua data from guaci_db.json, and computing mutual/changed hexagrams plus body-use and five-element relations. There is no logic that maps a hexagram to a guessed hidden object, so the advertised primary purpose is not implemented. Additionally, the code does rely on a database lookup, contradicting the '不查表' claim. This is a substantive description-behavior mismatch rather than a minor implementation detail.
The entire skill description, usage examples, and invocation phrases are presented only in Chinese, with no indication that other languages are supported or that Chinese is an intentional, justified region-specific limitation. Under the policy for natural-language violations, forcing a specific language without user opt-in is a reportable issue.
Without declared permissions the skill's intent is opaque and cannot be validated.
The trigger phrases include broad everyday language such as '起卦', '占卜', and especially '猜一猜', which can cause accidental activation during unrelated conversation. Overly broad activation increases the chance that the skill hijacks normal user interactions, leading to prompt-routing confusion, unintended behavior, or social-engineering opportunities if the skill inserts persuasive or off-task content.
The trigger list contains broad everyday phrases such as “算一算”, “测算”, “预测”, and “猜一猜”, which can match routine conversation and cause unintended skill activation. In a divination-style skill, accidental invocation can lead to surprise context capture, irrelevant responses, and policy/safety boundary confusion, especially because many triggers overlap with generic advice-seeking language.
The visible documentation labels the page as a 正易心法 skill, which suggests the direct-reading approach described in the manifest. However, the implementation depends on hardcoded 64卦 and 纳甲 tables and exposes different methods, creating an active contradiction between the stated intent and the code’s actual approach.
The manifest describes a skill for 射覆 using 《正易心法》卦脉法 to infer a hidden object directly from six-line structure, emphasizing no lookup tables or口诀. This file instead presents and implements two different divination systems—梅花易数 and 六爻纳甲—and uses substantial hexagram and 纳甲 tables, which is a semantic expansion beyond the claimed purpose.
The file is entirely written in Chinese and does not indicate any language selection, fallback, or user opt-in. In a user-facing skill, this can exclude users who do not read Chinese, cause misunderstanding of outputs, and weaken informed use of the skill, though it is an accessibility/product safety issue rather than a direct exploit primitive.
The manifest describes a skill for 射覆 that reads the qi tendency directly from a six-line structure to infer a hidden object, explicitly emphasizing interpretation rather than lookup formulas. This file instead performs random hexagram generation via 大衍筮法, loads a gua database, and attaches current ganzhi/calendar context, which is a broader divination-generation capability rather than direct 射覆 inference from an already-given hexagram structure.
The file presents all instructions, examples, and output guidance exclusively in Chinese, but does not indicate that the user can choose another language or that the skill is intentionally limited to Chinese-speaking users. Under the policy, forcing a specific language without opt-in is a natural-language policy concern.
The skill name and primary description are presented in Chinese, while English appears only in secondary fields, which may indicate a default language choice imposed by the skill metadata. There is no explicit statement offering users a language or locale preference, so the configuration may violate a language-choice policy if the platform expects user-controlled locale behavior.
The document declares lang="zh-CN", which hard-codes a specific language/locale for the skill UI. The file does not provide any user opt-in, alternate locale option, or justification that the skill is intended only for a China-specific compliance or regional context.
The code checks whether stdout is UTF-8 and unconditionally reconfigures it to UTF-8. This is a natural-language/locale behavior baked into the skill without offering the user a choice, which matches the policy concern about forcing a specific locale or language behavior.
The module docstring and later user-facing output indicate the skill is designed to operate in Chinese only, with no visible option for users to select another language or opt in to the locale. This is a natural-language policy concern because the skill forces a specific language in its interface and output.
All human-readable output strings in the formatter are hard-coded in Chinese, and the code exposes no mechanism for user language selection. This creates a language policy issue because the skill enforces a specific locale in normal operation.
The argparse description and help strings are entirely in Chinese, which fixes the command-line experience to one language without presenting a choice. Under the policy, locale constraints should either be optional for the user or clearly justified as region-specific.
The command-line description, help text, result keys, and printed output are all fixed in Chinese, with no option for users to select another language or locale. This is a natural-language policy concern because it imposes a specific language on all users without opt-in.
This code's docstrings, CLI usage text, JSON field names, and output strings are all hard-coded in Chinese, and there is no visible option for users to select another language or opt in to this locale. Under the natural-language policy rule, forcing a specific language without user choice is a policy concern even when the functionality itself is benign.
No suspicious patterns detected.
