Security audit
Security checks for vulnerabilities and agentic risk
This is a documentation-only workflow helper with no executable code or persistence, though its activation wording is broader than ideal.
This skill is reasonable to install if you want a planning/checklist helper for skill-vetting workflows. Be aware that it may auto-activate too broadly because of generic trigger terms like security, GitHub, first, and bug fix; use explicit invocation or narrow its triggers if your environment supports that.
The trigger sentence is written as a broad, ordinary user request pattern ('Help me ...'), which can cause the skill to activate during normal conversations that merely mention related topics. In an agent ecosystem, this creates prompt-scope confusion and unintended invocation risk, potentially steering workflows or outputs without clear user intent to use this specific skill.
This second trigger uses another generic natural-language request form ('I need a practical workflow ...') that is likely to overlap with ordinary user queries. Such overlap increases the chance of accidental activation, especially in systems that match skills based on approximate phrasing or keyword heuristics.
The trigger phrases are broad enough to match common, generic user requests such as security help, GitHub help, bug fixing, or general workflow support. This can cause the skill to activate outside its intended scope, increasing the chance that users are funneled into an unrelated or less appropriate workflow and potentially exposing them to unsafe or misleading guidance in security-sensitive contexts.
The skill description uses very broad trigger language such as 'security' and 'first' in a generic 'Use when' clause, which can cause the skill to activate for many unrelated conversations. Over-broad activation increases the chance that this skill intercepts user requests outside its intended scope, leading to misrouting, prompt-surface expansion, and unintended influence over higher-risk security-related tasks.
The keyword list includes vague everyday terms like 'security', 'first', 'before', and 'github' without any contextual guardrails. These common terms are likely to appear in benign user prompts, so the skill may spuriously trigger and inject its workflow into unrelated sessions, which is especially risky because it presents itself as a vetting/security helper and could override more appropriate handling.
The trigger keywords are excessively broad and include common terms such as "security", "first", "before", and "github", which can cause the skill to activate during ordinary unrelated conversations. In an agent system, this creates prompt-routing confusion and can unexpectedly inject this skill's workflow into tasks where it was not intended, increasing the chance of incorrect actions, disclosure of context, or unsafe automation decisions.
The trigger sentences are broad and generic enough to match ordinary requests containing words like "help me," "security," or "github," which can cause the skill to activate outside its intended scope. In an agent-routing context, overbroad activation can misroute user requests, override more appropriate skills, and increase the chance that sensitive or security-related tasks are handled by an unintended workflow.
These lines describe language-specific documentation variants but do not state that the user may choose their preferred language or locale. Under the policy, forcing or implying a fixed language without explicit opt-in can be a natural-language policy concern.
The example trigger sentences are malformed, generic, and lack clear invocation boundaries, so they do not teach a safe or narrow activation pattern. Without positive and negative examples, implementers may rely on partial phrase matching and unintentionally activate the skill for loosely similar requests, broadening exposure and causing workflow confusion.
The activation description says to use the skill for broad categories like work-productivity, security, or practical workflow support, but it does not clearly define scope limits or exclusion criteria. That ambiguity can cause over-activation and misrouting, especially because the skill is framed as generally helpful for analysis and implementation support, making it easier for unrelated requests to be captured.
The default prompt is overly broad and generic, combining common terms like 'help me' with a wide set of topics such as productivity, security, workflow, checklist, and implementation support. This can cause the skill to be selected in unintended contexts, increasing the chance of prompt hijacking through over-invocation, incorrect routing, or accidental access to sensitive user workflows where the skill is not the best match.
Enabling implicit invocation without clear constraints allows the platform to auto-select this skill based on weak or ambiguous matches. In this skill, the broad work-productivity/security positioning makes that especially risky because it can silently insert the skill into many conversations, causing unintended instruction influence, workflow confusion, or inappropriate handling of sensitive requests.
