Security audit
Security checks for vulnerabilities and agentic risk
This is a documentation-only workflow helper with overly broad activation wording, but no evidence of hidden code, data access, persistence, or destructive behavior.
Review the broad trigger behavior before installing, especially if your agent auto-selects skills. Prefer explicit invocation of this skill for Humanizer/generated-text workflow tasks, and avoid relying on its generic writing or bug-fix triggers.
The trigger phrases are broad, natural-language prompts that could match ordinary user requests unrelated to this specific skill. That increases the chance of accidental invocation, causing the agent to apply the wrong workflow, produce irrelevant outputs, or bypass more appropriate skills or safeguards through prompt-routing confusion.
The manifest description uses a catch-all activation condition such as 'or needs a practical workflow, artifact, checklist, analysis, or implementation support,' which is broad enough to match a large share of normal user requests. This weakens skill isolation and can cause inappropriate invocation, making downstream behavior less predictable and expanding the attack surface for adversarial prompt content embedded in skill instructions.
The trigger keywords are extremely broad and include generic terms like 'writing', 'text', 'editing', 'reviewing', and 'bug fix', which can cause this skill to activate for many unrelated requests. Over-broad activation increases the chance that the agent applies the wrong workflow, leading to prompt-routing errors, unintended instruction injection surface, and reduced reliability in security-sensitive contexts.
The trigger keywords are overly broad and include common terms like writing, text, editing, reviewing, and bug fix, which can match many unrelated everyday requests. This can cause the skill to activate outside its intended scope, increasing the chance of inappropriate routing, user confusion, and unintended application of workflow guidance in contexts where it does not belong.
The trigger sentence is overly broad and can cause the skill to activate on common, non-specific user phrasing rather than a clearly scoped request. In an agent environment, this increases the chance of misrouting unrelated tasks into this skill, producing inappropriate automation or bypassing more suitable safeguards and workflows.
The trigger sentence is extremely broad and written like ordinary user language, which increases the chance the skill will activate unintentionally during unrelated conversations. In an agent environment, accidental invocation can route tasks through the wrong workflow, causing confusing behavior, unexpected actions, or inappropriate handling of user data.
This trigger remains generic enough that many normal requests for a 'practical workflow' could match, creating ambiguous routing and unintended skill selection. Because the skill concerns broad productivity/editing tasks, the vague activation language makes over-triggering more likely rather than less.
This file is presented as a Chinese README, but key descriptive and activation content is written in English, which can effectively impose a language expectation on users without opt-in. The document does not state that bilingual content is intentional or offer a language/locale choice for activation and usage.
The example trigger sentences demonstrate vague invocation patterns that mirror the same ambiguous language in the description rather than providing concrete, bounded examples. This normalizes overly permissive routing behavior and makes it more likely that operators or automated systems will invoke the skill for loosely related tasks.
The example trigger phrases are phrased as broad natural-language requests that resemble ordinary user prompts rather than precise invocation patterns. This makes accidental activation more likely and reinforces ambiguous matching behavior, which can reduce routing reliability and cause the skill to intercept unrelated requests.
The default_prompt is extremely broad and ambiguous, combining generic phrases like 'help me' with a long, loosely scoped capability description. This can cause the skill to be invoked in situations the user did not clearly intend, increasing the chance of prompt hijacking, incorrect delegation, or unauthorized handling of unrelated tasks.
Enabling allow_implicit_invocation without clear trigger constraints allows the platform to auto-select this skill based on vague semantic matches. Because the skill description and prompt are broad, this raises the risk of unintended invocation, causing the skill to act on user requests outside its safe or intended scope.
Although additional context is present, the trigger remains ambiguous because it still matches generic requests for 'practical workflow' help without enough distinguishing constraints. This can cause accidental invocation across unrelated productivity or editing requests, reducing reliability and potentially exposing users to irrelevant or unsafe actions generated by the wrong skill.
The document explicitly labels separate English and Chinese instruction files, but does not state how the user chooses between languages or that language should follow user preference. This can create an implicit language/locale constraint without opt-in.
