Security audit
Security checks for vulnerabilities and agentic risk
This Markdown-only skill appears purpose-aligned overall, but it asks for broad execution/write authority while giving overly broad triggers and under-scoped state-changing/API-key workflows.
Review before installing. Use this only for explicit knowledge-exchange tasks, confirm any create/modify/delete/reset/import action before running it, avoid sharing sensitive knowledge content, and keep API keys out of logs, prompts, shell history, and untrusted endpoints.
The description says the free version '强化中文本地化' (strengthens Chinese localization), which indicates a language/locale constraint, but the file does not offer an opt-in choice or explain that the skill is intentionally limited to a Chinese-speaking context. Under the policy, forcing a specific language without user choice can be a natural-language policy violation.
The skill metadata and narrative say the free edition only supports basic exchange and single proposals, but the implementation section advertises broader create/query/modify/delete, import, and reset capabilities. This mismatch can cause an agent or user to invoke unintended state-changing operations, increasing the chance of unauthorized or destructive actions under a misleadingly limited description.
The skill describes modify, delete, reset, and import style operations without prominent warnings, confirmation requirements, or discussion of irreversible effects. In a tool that has write and exec capabilities, missing guardrails around destructive operations raises the risk of accidental data loss or unintended changes to exchange state.
The skill walks users through API-key registration and authenticated network transmission, including sending agent identifiers and bearer tokens, but lacks a dedicated privacy and security warning section. Without explicit handling guidance, users or agents may expose secrets in logs, shell history, prompts, or to untrusted endpoints.
The skill explicitly instructs external HTTP(S) transmission of registration data and later authenticated exchange traffic. External transmission is expected for this skill's purpose, but it remains security-relevant because it sends identifying information and uses API keys over the network; combined with broad triggers and exec capability, this could lead to unintended data exfiltration or secret misuse.
首次使用时注册 Agent 并获取 API 密钥.
# 注册获取 API 密钥（无需认证）
curl -X POST https://exchange.example.com/auth/register \
-H "Content-Type: application/json" \
-d '{"agentName": "我的Agent", "agentUrl": "https://my-agent.example.com"}'
# ...
The trigger condition is extremely broad, covering generic AI model use, intelligent dialogue, agent orchestration, and LLM applications. In an agent ecosystem this can cause the skill to activate in unrelated contexts, leading to unnecessary API-key use, external requests, or exchange actions when the user did not specifically request this tool.