AD agent-reach
Use the internet: search, read, and interact with 13+ platforms including Twitter/X, Reddit, YouTube, GitHub, Bilibili, XiaoHongShu (小红书), Douyin (抖音), WeChat Articles (微信公众号), LinkedIn, Boss直聘, RSS, Exa web search, and any web page. Use when: (1) user asks to search or read any of these platforms, (2) user shares a URL from any supported platform, (3) user asks to search the web, find information online, or research a topic, (4) user asks to post, comment, or interact on supported platforms, (5) user asks to configure or set up a platform channel. Triggers: "搜推特", "搜小红书", "看视频", "搜一下", "上网搜", "帮我查", "全网搜索", "search twitter", "read tweet", "youtube transcript", "search reddit", "read this link", "看这个链接", "B站", "bilibili", "抖音视频", "微信文章", "公众号", "LinkedIn", "GitHub issue", "RSS", "search online", "web search", "find information", "research", "帮我配", "configure twitter", "configure proxy", "帮我安装".
As a process D 41/100 · Unfinished process — weak spots: result and completion, when it triggers, inputs and preconditions
The same skill appears in 1 more place: ClawHub
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 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
- 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 · 1
✓ No critical or high findings
Medium and low: 1
-
medium Exfiltration
intent-browser-credential-storeSKILL.md:68Accesses a browser credential / cookie store (quoted — discussed, not commanded)> Server IPs may get 412. Use `--cookies-from-browser chrome` or configure proxy.
quoted
Files scanned: 2. Evidence is masked. Grey chips explain why severity was lowered.
Against the Agent Skills spec
✓ No remarks against the Agent Skills spec
Process rating: all ten parameters 41/100
- 0Result and completion. Does not say what the result is
- 0Inputs and preconditions. Does not say what the process needs to start
- 0Progress reporting. Says nothing while it works
- 20When it triggers. No condition that starts the skill
- 30Running it twice. 1 mutating operations with no state check
- 40Consistency. Frontmatter name (agent-reach) differs from the folder (skill-11)
- 50Steps. 2 steps
- 60Tools and files. Uses tools (bash, web, node) that frontmatter does not declare
- 100Failures and branches. 1 branches, has a failure section
- 100Execution cost. Instruction body is 1090 tokens
- low 16 top-level sections: this looks like several domains in one skill
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
- +4Description does not say when NOT to use the skill (false activations)
- +3Description length 907: 120–800 characters recommended
- +3No numbered steps or checklist
- +3Output format is not stated: the model decides each time
- +1No license
- +2Single-language instructions
- +5Description quotes 14 example trigger phrases
- +4Structure: 17 headings
- +4Has examples (14 code blocks)
Quality base 70; lint remarks subtract, signals add up to 100. Result: 83.