Security audit
Security checks for vulnerabilities and agentic risk
This is a GitHub workflow helper with broad trigger wording, but its documented behavior is coherent, disclosed, and focused on safe repository work.
Before installing, be aware that the skill may activate on common GitHub-related requests. For any mutating GitHub action, make sure the agent has the correct repository, PR or issue number, branch, and permission scope, and explicitly approve posting comments, rerunning checks, or changing repository state.
The trigger sentence begins with a very broad phrase ('Help me') and then appends generic requirement text, which can cause the skill to activate on many unrelated user requests. Overbroad activation increases the chance of incorrect routing, unintended execution of GitHub-oriented guidance, and prompt-surface expansion when other tasks are misclassified into this skill.
The trigger examples "Automate a safe GitHub issue triage workflow for this repository" and "Create commands for rerunning failed checks and reporting the result" are generic GitHub-related requests and do not clearly constrain when this skill should activate versus a more general coding or DevOps skill. The README provides examples but no explicit trigger boundaries or negative examples, which increases the risk of unintended invocation.
The manifest says to use the skill when a user asks for 'GitHub, pull request, issue, review comments, CI' or related workflow support, which spans very common development requests without clearly distinguishing when this skill should or should not activate. The description does not provide exclusion conditions or negative examples, increasing the chance of unintended invocation for ordinary repository help requests.
The trigger keywords include generic terms like 'GitHub', 'pull request', 'issue', and 'CI', all of which commonly appear in everyday developer conversation. Because the file does not pair these with limiting conditions or negative examples, the trigger scope is too broad for reliable activation.
The skill description is broadly scoped around common GitHub activities and can match routine developer conversations without sufficient gating conditions. This can cause unintended skill invocation, pulling the agent into repository or automation workflows when the user did not explicitly request that behavior, increasing the chance of over-collection of context or unnecessary action suggestions.
The trigger list contains very common terms such as GitHub, issue, pull request, review comments, and CI, which appear in many normal engineering discussions. Without scope constraints, these high-frequency keywords can cause accidental activation and route unrelated conversations into this skill, potentially leading to irrelevant automation guidance or unsafe assumptions about repository context.
Although the second trigger adds a little context ('I need a practical workflow for ...'), it still embeds generic requirement prose instead of a precise invocation pattern. This ambiguity can still match loosely related requests and route users into a GitHub-interaction skill when they asked for something broader or different, creating reliability and safety issues through unintended tool or workflow selection.
This file is entirely presented in Simplified Chinese and does not indicate that language is optional or that users can choose another locale within this document. The policy requires flagging language or locale constraints when they are imposed without user opt-in or clear justification.
