Security audit
Security checks for vulnerabilities and agentic risk
This is a coherent Toutiao writing and local article-audit skill that discloses its local report files and shows no network, credential, account-login, or background behavior.
Use this skill for Chinese Toutiao article drafting and pre-publication review. When running the script, point it only at files or directories you intend to audit because it will read matching drafts and create local JSON report files there; its compliance output is guidance, not a guarantee that Toutiao will approve a post.
The skill advertises script-driven file read/write behavior (python scripts/toutiao_audit.py, reports written to user directories) but does not declare any explicit tool scope such as permissions or allowed-tools. That creates an authorization ambiguity where a host agent may grant broader filesystem access than users expect, increasing the risk of unintended file access or modification.
Broad trigger keywords can cause the skill to activate for generic content-writing requests outside the user's intent. In this skill, accidental activation matters because it can steer users toward local script-based auditing and file handling behavior, expanding the skill's effective reach beyond a narrowly scoped Toutiao workflow.
Generic trigger terms like broad article-creation phrases are unreliable for safe routing because they overlap with many ordinary writing tasks. This can lead to inappropriate invocation of the skill's prescriptive workflow and associated script usage in contexts where the user did not request Toutiao-specific review.
Ambiguous keywords that are not clearly bound to the target platform can misroute unrelated requests into this skill. While the content is primarily advisory rather than overtly dangerous, misactivation still increases confusion and could prompt unnecessary file-based auditing actions or irrelevant policy guidance.
The file is entirely framed in Chinese and presents the review standard as the required process, but it does not indicate that Chinese is optional or that users may choose another language/locale. Under the policy, forcing a specific language without user opt-in can be a natural-language policy violation unless the locale constraint is explicitly justified.
This markdown file is entirely written in Chinese and does not indicate that the skill supports other languages or that Chinese is an explicit user-selected option. Under the stated policy, forcing a specific language without user opt-in can be a natural-language policy violation unless the locale constraint is clearly documented and justified.
The manifest and module docstring describe a skill for content production quality control and auditing, which implies analysis of article files. However, the implementation also creates JSON report files on disk (audit_report.json or .audit.json), adding a file-modification behavior beyond pure auditing/inspection.
This code writes JSON audit results to disk in both batch and single-file modes via write_text, which is a file-modifying operation. Although the script prints the saved path after writing, there is no prior confirmation prompt, upfront warning in the module docstring, or explanatory comment near the write operations disclosing that running the script will create report files alongside the target path.
The file title and all checklist content are written entirely in Chinese, presenting the checklist as a required publication gate rather than an optional locale-specific variant. Under the policy, forcing a specific language without user opt-in is a natural-language locale violation unless the locale restriction is clearly justified, which is not stated here.
This markdown template appears to require use in Chinese by presenting all instructions, examples, and labels exclusively in Chinese. Under the policy rule for language or locale constraints, a skill should not force a specific language unless it offers user choice or clearly states a justified region-specific purpose.
No suspicious patterns detected.
