DF linkfox-os
linkfox-os — Cross-border e-commerce AI agent system with 6 specialized agents covering the full seller workflow: (1) General Assistant (default) — platform data queries across Amazon/TikTok Shop/eBay/Walmart/Shopee/Ozon, Keepa/SIF/SellerSprite analytics, Google Trends, 1688 sourcing, patent/IP search, PDF analysis, web search; (2) Market Analysis Agent (linkfox-market-analysis-agent) — 5-dimension market research (market overview, competitor analysis, review mining, keyword research, compliance detection), structured HTML reports; (3) Product Selection Agent (linkfox-product-selection-agent) — keyword-based selection, viral product prediction, condition-based filtering, benchmark selection across 7 platforms, risk assessment; (4) Listing Agent (linkfox-listing-agent) — Amazon Listing creation/optimization/scoring (benchmark, rewrite, create, batch modes), keyword matrix, compliance scan; (5) Image Agent (linkfox-image-agent) — product/cloth image collections, image fission, bestseller replication, mannequin-to-model, white background, scene, selling point, A+, model, close-up images; (6) Video Agent (linkfox-video-agent) — image-to-video, AI sales talking video, viral video replication. Use when: product selection, market analysis, competitor research, keyword research, review mining, Listing writing/optimization, product image generation, video generation, IP/patent detection, trend analysis, 1688 sourcing, cross-platform data queries, or any multi-step e-commerce workflow. **内置账号引导（onboarding）**：未配置 API key / 鉴权失败（401）/ 计费不足（402 或消息含'积分余额不足/余额不足/请充值/quota exceeded/insufficient balance'）时，自动进入引导流程——脚本化手机号注册取 key、列套餐、生成支付二维码，见 references/onboarding.md。用户说'没配 key/鉴权失败/积分不足/充值/recharge/注册/手机号注册'也触发。**用户素材上传（File Upload）**：用户先提供本地图片/文档/视频给下游 agent 使用时（触发短语：'我有一张参考图 / 帮我上传商品图 / 帮我传一份文档 / 附一份参考视频 / 用这张图生成…'），走 `<skill>/scripts/upload/upload_file.py` 拿到 `file://` 虚拟路径后塞进下一步 prompt，见 SKILL.md §14。
linkfox-os — Cross-border e-commerce AI agent system with 6 specialized agents covering the full seller workflow: (1) General Assistant (default) — platform…
As a process F 61/100 · Will not run — References files that are not bundled: /path/to/output/202607.../reviews-123.json, url, /absolute/path/to/filename.json
What is at stake
Medium-severity findings: the skill is probably honest, but read what alarmed the scanner.
Below is the worst case for this category. The finding here is medium: the guard saw a sign, not a proof.
The skill contains commands that delete files, rewrite disks or execute code fetched from the network. The agent may run them without asking if it believes the instructions require it.
Replace destructive commands with safe equivalents that ask for confirmation, scope them to one folder, and stop piping curl into a shell: pin a version and a checksum.
Below is the worst case for this category. The finding here is medium: the guard saw a sign, not a proof.
The files contain invisible characters, encoded commands or comments hidden from readers but visible to the model. What you read differs from what the agent sees.
Remove invisible characters (they usually sneak in through copy-paste) and encoded strings: no catalog will pass them. Instructions for the model must be readable by a human too.
Below is the worst case for this category. The finding here is medium: the guard saw a sign, not a proof.
The instructions or scripts send data out: environment variables, keys, file contents, chat history. You may never notice, because the agent performs the upload as "part of the task".
If the upload is not needed for the task, remove it: catalogs flag such skills and delist them. If it is needed, name the destination explicitly, say what leaves the machine, and give the user a switch.
How to improve
- Shorten the description to 1024 characters.
- The SKILL.md body is over 5,000 tokens: move reference detail into references/ and load it when needed.
- The text references files that are not there: add them or drop the references.
- Your own cases (evals/evals.json, 4–6 real requests with expected answers): the full check would then run those instead of a model-drafted suite.
- A spec.yaml with trigger phrases and assertions — a behaviour contract for CI; `skilltest init` writes a template.
Guard findings · 9
✓ No critical or high findings
Medium and low: 9
-
medium Dangerous commands
cmd-shell-rcreferences/onboarding.md:108Writes to a shell startup fileecho 'export LINKFOXAGENT_API_KEY="你的key"' >> ~/.zshrc # macOS 默认 zsh
-
medium Dangerous commands
cmd-shell-rcreferences/onboarding.md:110Writes to a shell startup fileecho 'export LINKFOXAGENT_API_KEY="你的key"' >> ~/.bashrc # Linux 默认 bash
-
medium Obfuscation
obf-base64-blobreferences/skills-third-platforms.md:18Long base64-looking blob| `linkfox-echotik-list-product` | TikTok Shop | TikTok Shop商品搜索+筛选（EchoTik）| `keyword`, `region`, `minT…Cnt`, `productSortField` | 商品列表JSON（title/price/tota…Cnt/tota…Amt/pr
-
medium Exfiltration
net-redirectable-api-keyscripts/linkfox_os.py:55Helper sends the API key to a host configured by an environment variable — the key can be redirected to another serverAPI key + configurable base URL from environment
-
medium Exfiltration
net-redirectable-api-keyscripts/onboarding/setup_common.py:35Helper sends the API key to a host configured by an environment variable — the key can be redirected to another serverAPI key + configurable base URL from environment
-
medium Exfiltration
net-redirectable-api-keyscripts/upload/upload_common.py:30Helper sends the API key to a host configured by an environment variable — the key can be redirected to another serverAPI key + configurable base URL from environment
-
low Secrets in code
secret-high-entropy-tokenreferences/onboarding-api.md:317High-entropy token-like string (may be an id, hash or a credential) (detector / deny-list definition; documentation table row)| Header | `source: ai-linkfox-web`、`Content-Type: application/json;charset=UTF-8`、`uid: eyJh…iJ9`（固定常量，env `LINKFOX_LOGIN_FIXED_UID` 可覆盖） |
detectortable -
low Secrets in code
secret-high-entropy-tokenreferences/onboarding-api.md:433High-entropy token-like string (may be an id, hash or a credential) (quoted — discussed, not commanded)示例：accessToken 的 JWT header `{"a":"6a22c8b0***16a","u":"2074679****56403968","alg":"HS256"}` + userId `207467****6403968` → uid = `base64({"a_id":"6a22c*****c916a","d_id":"20746794*****403968"})` = `equoted -
low Secrets in code
secret-high-entropy-tokenscripts/onboarding/setup_common.py:54High-entropy token-like string (may be an id, hash or a credential) (quoted — discussed, not commanded)_LOGIN_FIXED_UID = os.environ.get("LINKFOX_LOGIN_FIXED_UID") or "eyJh…iJ9"quoted
Files scanned: 25. Evidence is masked. Grey chips explain why severity was lowered.
Against the Agent Skills spec
- error
description-longdescription is 1924 chars, limit 1024 - warning
frontmatter-yamlSKILL.md: the frontmatter is not valid YAML (YAML parse error: Nested mappings are not allowed in compact mappings at line 2, column 14: description: linkfox-os — Cross-border e-commerce AI agent system with 6 specia… ^ ); fields were read line by line. The usual cause is a colon inside an unquoted value - warning
body-longSKILL.md body ≈ 7463 tokens (recommended < 5000); move details to references/ - warning
missing-refreference to a missing file: /path/to/output/202607.../reviews-123.json - warning
missing-refreference to a missing file: url - warning
missing-refreference to a missing file: /absolute/path/to/filename.json - warning
missing-refreference to a missing file: /Users/xxx/project/.linkfox-os/output/202607161030/amazon-search-results.json - warning
missing-refreference to a missing file: /local/path/filename.json - warning
missing-refreference to a missing file: /abs/path/filename.json - warning
missing-refreference to a missing file: <url> - warning
missing-refreference to a missing file: /path - warning
missing-refreference to a missing file: references/skills-*.md - note
description-budgetdescription takes 1924 of the ~15000-char shared budget for all skills
Process rating: all ten parameters 61/100
- 0Tools and files. 9 referenced file(s) missing: /path/to/output/202607.../reviews-123.json, url, /absolute/path/to/filename.json
- 0Inputs and preconditions. Does not say what the process needs to start
- 30Running it twice. 12 mutating operations with no state check
- 60Result and completion. Output format stated, no completion criterion
- 70Execution cost. Instruction body is 7463 tokens
- 100Steps. 93 steps
- 100When it triggers. States when to use and when not to
- 100Failures and branches. 2 branches, has a failure section
- 100Consistency. Name and required fields are in place
- 100Progress reporting. Reports progress
- low 16 top-level sections: this looks like several domains in one skill
- low The response is described with custom markup (27 tags): a typed call is more reliable
Everything here is measured from the skill text rather than judged by a model, so the numbers are checkable. A parameter weighs more when it is a more common reason for the process to stall.
Quality signals
- +5Description has no quoted example phrases that should trigger the skill
- +4Description does not say when NOT to use the skill (false activations)
- +3Description length 1924: 120–800 characters recommended
- -5TODO / placeholder text left in the skill
- -4Absolute local paths (C:\Users, /home/…): not portable
- +1No license
- +2Single-language instructions
- +4Structure: 50 headings
- +3Step-by-step instructions: 93 items
- +3Output format is stated explicitly
- +4Has examples (30 code blocks)
- +4Reference files are cited in the instructions (12 of 12)
- +3All 1 scripts are documented
Quality base 70; lint remarks subtract, signals add up to 100. Result: 22.